Journal of Emerging Technologies in Web Intelligence Contents

Journal of Emerging Technologies in Web IntelligenceISSN 1798-0461Volume 2, Number 2, May 2010Special Issue: Emerging Technologies in Internet Security and NetworksGuest Editors: Himanshu Aggarwal and Vishal GoyalContentsGuest EditorialHimanshu Aggarwal and Vishal Goyal79SPECIAL ISSUE PAPERSPerformance of Dominating and Adaptive Partial Dominating Sets in AODV Routing Protocol forMANETsA. Nagaraju, S. Ramachandram, and B. EswarBee-Inspired Routing Protocols for Mobile Ad Hoc NetworkDeepika ChaudharyDistance and Frequency based Route Stability Estimation in Mobile Adhoc NetworksAjay Koul, R. B. Patel, and V. K. BhatInvestigation of Blackhole Attack on AODV in MANETAnu Bala, Raj Kumari, and Jagpreet SinghCopyright Protection of Gray Scale Images by Watermarking Technique using (N,N) Secret SharingSchemeSushma Yalamanchili and M. Kameswara RaoBroadband Integrated Services over Proposed Open CPE ArchitectureKushal RoyRecovery Based Architecture to Protect Hids Log Files using Time StampsSurinder Singh Khurana, Divya Bansal , and Sanjeev SofatSoftware RadioVarun Sharma and Yadvinder Singh MannSDR and Error Correction using Convolution Encoding with Viterbi DecodingShriram K. Vasudevan, Siva Janakiraman, and Subashri VasudevanWebinar – Education through Digital CollaborationAnuradha Verma and Anoop SinghEfficient Visual CryptographySupriya A. Kinger80868996101106110115122131137

Multistage Interconnection Networks: a Transition from Electronic to OpticalRinkle Rani Aggarwal, Lakhwinder Kaur, and Himanshu AggarwalWeb Based Hindi to Punjabi Machine Translation SystemVishal Goyal and Gurpreet Singh LehalProtecting Data from the Cyber Theft – a Virulent DiseaseS. N. Panda and Vikram Mangla142148152

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 79Special Issue on Emerging Technologies in Internet Security and NetworksGuest EditorialComputer systems have been the core of information technology to provide signal processing and storage capabilitiesand form the basis for the Internet connectivity. Computer systems act as the fundamental core hardware forimplementation of the physical layer of today's worldwide information super highways. Networking technologiesprovide end-users with the facilities for network connection, data access, and communications through wired andwireless IP-based networks. On the other hand, communication technologies provide means for signal transmission andreception via efficient multiple access capability through various media, including copper wires, optical fibers, radiofrequency (RF) and other approaches (such as infrared, laser, etc.). Integration of computer systems, networks, andcommunications is an important trend for future information technologies (ITs).With the advances in the communication and internet technologies the Information security, data integrity and safeexchange of data and Information are most important and contemporary issues in the today’s world. With a mission toreport emerging a state-of-the-art research in Technologies in Network security and Computer Networks, the mostimportant areas of information technology, a special issue of JEWTI has been dedicated The research contributionshave been collected through the National Conference –Cum- Workshop on Information Security and Network (ISAN).The mission of ISAN was to promote research, development and new inventions of Information Security and Networksincluding mobile and wireless technologies DCAS, intrusion detection etc.This Issue of the Journal is designed for researchers, developers, practitioners, policy-makers, professional trainers,educators, and other specialists.Guest Editors:Dr. Himanshu Aggarwal, Punjabi University, India.Dr. Vishal Goyal, Punjabi University, India.Himanshu Aggarwal, Ph.D., is Reader in Computer Engineering at Punjabi University, Patiala. Hehas more than 15 years of teaching experience and served academic institutions such as Thapar Instituteof Engineering & Technology, Patiala, Guru Nanak Dev Engineering College, Ludhiana and TechnicalTeacher’s Training Institute, Chandigarh. He is an active researcher who has supervised many M.Tech.Dissertations and contributed 40 articles in various Research Journals and Conferences. He is on theeditorial Board of several journals of repute. His areas of interest are Information Systems, ERP andParallel Computing. Himanshu Aggarwal can be contacted at: himagrawal@rediffmail.com.Vishal Goyal, MCA, M.Tech., is senior lecturer at Department of Computer Science, PunjabiUniversity Patiala, Punjab, India. He has more than 10 years of teaching, research and software industryexperience. He has about 50 publications in various journals and conferences of national and internationalrepute. He is member of editorial board of various journals of computer science of national andinternational repute. He is Technical team member of E&R Department of Infosys Technologies. DST hasfunded him research project worth Rs 28Lakhs for development of Punjabi Grammar checker. He is veryactive researcher in the field of Natural Language Processing. He has guided approx. 25 M.tech. studentsand 15 M.Phil Students for their research work. He has been honored with Young Scientist award byPunjab Academy of Sciences in 2005. He is being regularly invited by colleges and universities for hisexpert talks. He can be reached at vishal.pup@gmail.com or http://www.punjabiuniversity.ac.in.© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.79-79

80 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Performance of Dominating and Adaptive PartialDominating Sets in AODV Routing Protocol forMANETsA. Nagaraju *, Dr. S. Ramachandram**, B. Eswar**** Head & Associate Prof in I.T , Kamala Institute of Tech & Science, Huzurabad, Karimnagar, A.P** Head & Prof in CSE , Osmania University , Hyderabad. A.P, INDIA***Assistant Prof in I.T, Kamala Institute of Tech & Science, Huzurabad, Karimnagar, A.P. INDIAAbstract— A mobile ad hoc network (MANET) is a wirelessnetwork that does not rely on any fixed infrastructure (i.e.,routing facilities, such as wired networks and access points),and whose nodes must coordinate among themselves todetermine connectivity and routing. The broadcast can targeta portion of the network (e.g. gathering neighborhoodinformation), or the entire network (e.g., discovering routes ondemand). Broadcasting of signaling and data in MANETs raiseredundant transmission of control packets to overcome theseproblems we applied dominating set and Adaptive partialDominating (APDP) approach to existing routing protocolssuch as Ad-hoc On-demand Distance Vector (AODV). Thefocus of this paper is to apply the concept of DS and APDP toAODV and evaluate the performance of dominating sets inAODV that improve broadcasting, End-to-End Delay, Networkload, Packet Latency, and also maintains secure packettransmission.IndexTerms—Adaptive partial dominating, Dominating sets,AODVI. INTRODUCTIONMobile IP and wireless networks accessing the fixednetworks have provided support for the mobility. But it isstill restrictive in forcing the connectivity at least to thecore network. It puts impediments on supporting the truemobility in the network. In this connection, one areawhich is getting much attention in last couple of years isMobile Ad Hoc Networks (MANETs). A MANET(Mobile Ad-Hoc Network) is a type of ad-hoc networkwith rapidly changing topology. Since the nodes in aMANET are highly mobile, the topology changesfrequently and the nodes are dynamically connected in anarbitrary manner.As defined in [1] a MANET is an autonomoussystem of mobile nodes. MANET nodes are equippedwith wireless transmitters and receivers using antennawhich may be omni directional, highly-point-to-point,possibly steerable, etc. At a given point in time,depending on the positions of the nodes and theirtransmitters and receivers coverage patterns,transmission power levels and co-channel interferencelevels, a wireless connectivity in the form of a random,multi hop graph or ad hoc network exists among thenodes. This ad hoc topology may change with time as thenodes move or adjust their transmission and receptionparameters. There is current and future need for dynamicad hoc networking technology. The emerging field ofmobile and nomadic computing, with its currentemphasis on mobile IP operation, should graduallybroaden and require highly-adaptive mobile networkingtechnology to effectively manage multi hop, ad-hocnetwork clusters which can operate autonomously or,more than likely, be attached at some point(s) to thefixed Internet.There are two broad categories of unicast routingprotocols for MANETs, proactive and reactive. Withproactive routing (e.g., OLSR [21]), nodes keep routinginformation to all nodes in the network, not subject to anyexisting data flow. OLSR is a link state protocol using anoptimized broadcast mechanism for the dissemination oflink state information. In reactive routing (e.g., AODV[3]), routes are found on demand and nodes find routes totheir destinations as they are needed. Route discoverystarts by broadcasting a route request (RREQ) messagethroughout the network. This message is relayed until itreaches a node with a valid route to the destination, or thedestination itself. Once this happens, a routereply (RREP)message is sent back to the source by reversing the pathtraversed by the RREQ message. Only after receiving thecorresponding RREP message can the source startsending packets to the destination. Reactive and proactiverouting can be combined, resulting in hybrid protocols(e.g., ZRP 20]). In this case, routes to some nodes(usually the nearest ones) are kept proactively, whileroutes to the remaining nodes are found on-demand.Neighbor-knowledge-based methods mainlydepend on the following idea: To avoid flooding thewhole network, a small set of forward nodes is selectedsuch that the forward node set forms a connecteddominating set (CDS). A node set is a connecteddominating set if every node in the network is either inthat set or the neighbor of a node in that set. Then thechallenge is to select a small set of forward nodes in theabsence of global network informationIn our proposed paper we focused on applyingAPD to AODV and to evaluate the performance ofDominating sets in Ad-hoc On Demand Distance VectorRouting algorithm(AODV) that improve broadcasting, Endto-EndDelay, Network load, Packet Latency, and alsoMaintains secure packet transmission . To do this, conceptsfrom domination in graphs have been explored (i.e.© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.80-85

82 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010messages are handled may differ among differentprotocols, but their functionality remains the same: arequest is relayed until it reaches a node with a valid routeto the destination or the destination itself, which triggers areply message sent back to the originator. Severalparameters (such as how long to keep requests in a cache,timeouts for requests, timeouts for hellos) are subject totuning, and the choices made may result in improvementsin the protocol performance. However, RREQs arepropagated using either an unrestricted broadcast or anexpanding ring search. In either case, the resultingflooding operation causes considerable collisions ofpackets in wireless networks using contention-basedchannel access.In addition to applying DP to reduce the number ofnodes that need to propagate RREQs transmitted onbroadcast mode, information regarding prior routes to adestination is used to unicast RREQs to a region close tothe intended destination, so that broadcast RREQs arepostponed as much as possible and occur (if necessary)only close to the destination, rather than on network-widebasis. This RREQ Algoritm presents the pseudo-code forthe modified RREQ. A route request (RREQ) is handledas follows:• If the source of a RREQ does not have any previousknowledge about the route to the destination or is retryingthe RREQ, it calculates its forwarder list using DP, andbroadcasts the packet (Lines 8, 9, and 14).• On the other hand, if the source of a RREQ hasknowledge about a recently expired route to thedestination, and there is a valid route to the next hoptowards the destination (Lines 2, 3, and 4), the nodecalculates the forwarder list using DP (Line 9), but insteadof broadcasting the RREQ packet, the node unicasts thepacket to the last known next hop towards the destination(Line 12).Upon receiving a route request, a forwarder thatcannot respond to this request calculates its ownforwarder list using the information provided in theRREQ packet (i.e., forwarder list, second to previousforwarder list, and source node) and broadcasts or unicaststhe packet (depending on which one of the two first casesapply) after updating it with its own forwarder list.RREQ AlgorithmData : n i, destination D, B i , U iResult : Unicast the RREQ, or Broadcast the RREQBegin1 if recently expired route to D and not retryingthen2 NextHop ← previous_nextHop(D)3 if validRoute(NextHop) then4 result←Unicast5 else6 result←Broadcast7 else8 result←Broadcast9 F i ←DP( n i, B i , U i )10 Update RREQ packet with F i11 if result== Unicast then12 Unicast the RREQ packet to NextHop13 else14 broadcast the RREQ packetendEventually, the RREQ reaches a node with a route tothe destination or the destination itself. Our approachattempts to reduce the delay of the route discovery byunicasting a RREQ towards the region where thedestination was previously located. The success of thisapproach depends on how fresh the previous known routeto the destination is, and how fast the destination node ismoving out of the previous known location. If anintermediate node has completely removed any route tothe destination, the RREQ is then broadcasted. Theintended effect is to postpone the broadcast of a RREQ tothe region closest to the destination. In the case that theunicast approach fails, or there is no previous route to thedestination, the source broadcasts by default.Because of topology changes, nodes may not havecorrect two-hop neighborhood information, which mayresult in forwarding lists that do not cover all nodes inthe neighborhood. However, this is not a major problemwhen the request is broadcasted, because a nodeincorrectly excluded from the forwarder list may alsoreceive the request and is able to respond in the case it hasa route to the destination.C. Enhanced Dominant Pruning AlgorithmIn their paper [6], wei Lou and Jie Wu proposed twoenhanced dominant pruning algorithms: the TotalDominant pruning (TDP) algorithm and ParatialDominant Pruning (PDP).In TDP algorithm, if node v can receive a packetpiggybacked with N(N(v)) from node u , the 2-hopneighbour set that needs to be covered by v’s forwardnode list F is reduced to U = N(N(v)) – N(N(u)). Themain objective of the TDP algorithm is that 2-hopneighbourhood information of each sender is piggybackedin the broadcast packet which results in consumption ofmore bandwidth.D. Partial Dominant Pruning Algorithm:Just like the DP algorithm, in PDP, noneighbourhood information of the sender is piggybackedwith the broadcast packet. Apart from excluding N (u)and N (v) from N (N (v)) as in the DP algorithm, we canhere exclude some more nodes from neighbours of eachnode in N (u) ∩ N (v). Such a node set is denoted by P(u,v) (or simply P) = N(N(u) ∩ N(v)). Then 2-hopneighbour set U in the PDP algorithm can be given by U= N(N(v)) – N(u) – N(v) – P since P is a subset ofN(N(u)), we can easily see P can be excluded fromN(N(v)). Also it can be proved that U is subset of N(B),when P = N(N(u)∩N(v)), U = N(N(v)) – N(u) – N(v)-Pand B = N(v) – N(u).E Adaptive Partial Dominant Pruning Algorithm[19]Adaptive dominant pruning algorithm (APDP) issimilar to PDP. However, besides excluding N (u), N (v)and P from N (N (v)) as mentioned in PDP algorithm,adjacent nodes of U are eliminated from U.© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 83APDP Algorithm1 Node v uses N(N(v)) ,N(u), and N(v) to obtain P= N (N(u) ∩ N(v) ), U 1 = U – E where U =N(N(v)) – N(u) – N(v) – P and E is the set ofequivalent and adjacent nodes in U and B =N(v) – N(u)2 Node v calls the selection process to determineF.Now consider Fig 1 in the example of sample ad-hocnetwork with 12 nodes. Table 1 shows each node in Fig 11-hop and 2-hop neighbor nodes. Here we illustrate thedifference between PDP and APDP algorithms.Table 2: PDP algorithmu v P U B FØ 6 Ø 1,3,4,8,10,11 2,5,7,9 7,2,96 7 1,3,6,7 10,12 4,8,11 116 2 2,4,6,8,11 Ø 1,3 [ ]6 9 1,6,9 9 10 107 11 Ø 9 10,12 109 10 Ø 7,12 11 11Table 3: APDP algorithmU V P U B FØ 6 Ø 1,4,8,11 2,5,7,9 7,26 7 1,3,6,7 10,12 4,8,11 116 2 2,4,6,8,11 Ø 1,3 [ ]6 9 1,6,9 9 10 107 11 Ø 9 10,12 109 10 Ø 7,12 11 11The lower bound as per the AMCDS (ApproximationMinimum Connected Dominating Set) reduces theminimum connected dominating set to {2, 6, 7, 11}minimizing the number of forward nodes to 4. Accordingto our proposed model number of total forwarding nodesincluding source node is 5 i.e. {2, 6, 7, 10, 11} where as itis 6 in PDP.Fig 1 An Ad-hoc Network with 12 nodes.Table 1: Two hop neighbors of each nodeV N(v) N(N(v))1 1,2,5 1,2,3,5,6,7,92 1,2,3,6,7 1,2,3,4,5,6,7,8,9,113 2,3,4 1,2,3,4,6,7,84 3,4,7,8 2,3,4,6,7,8,11,125 1,5,6,9 1,2,5,6,7,9,106 2,5,6,7,9 1,2,3,4,5,6,7,8,9,10,117 2,4,6,7,8,11 1,2,3,4,5,6,7,8,9,10,11,128 4,7,8,12 2,3,4,6,7,8,11,129 5,6,9,10 1,2,5,6,7,9,10,1110 9,10,11 5,6,7,9,10,11,1211 7,10,11,12 2,4,6,7,8,9,10,11,1212 8,11,12 4,7,8,10,11,12For PDP algorithm, node 6 again has same forwardnode list F (Ø,6) = [7,2,9]. From P(6,7) = {1,3,6,7 }, wehave U(6,7) = N(N(7)) – N(6) – N(7) – P(6,7) = { 10,12 }.The forward node list for 7 is F(6,7) = {11 }, Similarly ,from P(6,2)= { 2,4,6,8,11} , we have U(6,2) = N(N(2)) –N(6) –N(2) –P(2,6) = Ø and , then , F(6,2) = {10 }.Therefore, the total number of forward nodes is 1+3+2 =6. The details of P, U, B and F are represented in thefollowing Table 2.The total number of forwarding nodes according toPDP is in the give example is 6 including source node i.e.{ 6,2,7,9,11,10 }.In our proposed model APDP, an enhanced version ofPDP, the definition of existing U has been broadened to anew U to check and exclude if it contains any adjacentnodes. The results of the proposed model are presented inTable 3IV. SIMULATIONS RESULTS AND SIMULATIONPARAMETERSWe used the simulator glomosim-2.03,[8] to run thesimulation Table 3 summarizes the simulation parameterswe used. The simulation time was 15 minutes accordingto simulator clock. A total of 45 nodes were randomlyplaced in field of 500 X 500 m 2 and in field of 2000 X2000 m 2 . Power range of each node is 250m. Weperformed the simulation for AODV with DP algorithm.Table 3 Simulation ParametersParameter Value DescriptionNumber of nodes 160 and 40 Simulation NodesField range x 500m and 2000 X-DimensionField range y 500m and 2000 Y-DimensionPower range 250m Nodes power rangeMac protocol IEEE 802.11 MAC layer protocolNetwork Protocol AODV &RoughAODVTransport Layer UDPProtocolPropagationfunctionFREE-SpaceNetwork LayerTransport LayerPropagationFunctionNode placement Random Nodes aredistributed inrandom mannerSimulation time 15M According tosimulation clockMobility Interval 10-30sec Pause time of node© 2010 ACADEMY PUBLISHER

84 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010We run each simulation four times with different nodepause time varying from 15 to 30sec with a step intervalof 5 sec. The graphs are presented in results section.ResultsThe performance of proposed protocol is evaluated usingthe following metrics:Packet delivery ratio (Throughput): Packet deliveryfraction is the ratio between the number of packetsoriginated by the application layer CBR sources and thenumber of packets received by the CBR sinks at the finaldestinations. Packet delivery Ratio is higher for AODVwith Dominating Sets than that of conventional AODV(figure 3 & 7 )Number of Control Packets: The total number ofcontrol packets occurred by different nodes is less inAODV with Dominating Sets than that of conventionalAODV(figure 4 & 8).Number of Route requests: It is the number of controlpackets generated by all the nodes in the simulation. Thenumber of Route Request packets is less in AODV withDominating Sets (figure 2 & 6 ) compare to conventionalAODV.End-to-End Delay: The End-to-End Delay of AODVwith Dominating Sets is better when compared withconventional AODV (figure 5 & 9).Performance of AODV with Dominatin sets andAODV in the Terrain Dimensions (500, 500)RouteThroughputs inRequests(persecondsnodes)Comparision Route Requests in AODV and AODV withDominating sets700600500400300200100020 40 60 80 100 120 140 160No of NodesAodvAodvdomFig 2 Route Request packet ComparisonComparision of Throughputs in AODV and AODV withDominating sets14000120001000080006000400020000AodvdomAodv20 40 60 80 100 120 140 160No of NodesAODVAODVDOMAODVAODVDOMEnd-to-End Delayin secondsComparision of End-to-End Delay in AODV and AODVwith Dominating sets0.050.040.030.020.010Aodv1 2 3 4 5 6 7 8No of NodesAodvdomFig 5 End-to-End Delay in SecAODVAODVDOMPerformance of AODV with dominating sets andAODV in the Terrain (2000, 2000) up to 40 nodesRouteRequestpackets sentThroughputs(inControl Packetsseconds)Comparing Route Request packets sent in AODV andAODV with dominating sets2000Aodv15001000500010 15 20 25 30 35 40No of NodesAodvdomFig 6 Route Request packet ComparisionComparision of Throughputs in AODV and AODV withdominating sets80007000Aodvdom60005000400030002000Aodv1000010 15 20 25 30 35 40No of nodes2000150010005000Fig 7 Throughput comparisonComparision of CTRL Packets in AODV and AODV withDominating setsAodvAodvdom10 15 20 25 30 35 40No of nodesFig 8.Control Packets ComparisonAODVAODVDOMAODVAODVDOMAODVAODVDOMControl Packets800600400200Fig 3 Throughput comparisonComparision of Control Packets in AODV and AODV withDominating sets020 40 60 80 100 120 140 160No of NodesAodvAodvdomFig4.Control Packets ComparisonAODVAODVDOMEnd-to-End delay insecondsComparision of End-to-End Delay in AODV and AODVwith Dominant sets0.350.3Aodv0.250.2AODV0.15AODVDOM0.1Aodvdom0.05010 15 20 25 30 35 40No of nodesFig 9 End-to-End Delay in Sec© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 85V. CONCLUSIONS AND FUTURE WORKWe presented an enhanced dominant pruning approachthat allows pruning redundant broadcasts even more thanthe conventional dominant pruning heuristic. Redundantbroadcasts increase the number of packet collisions, andconsequently delay the response for RREQs in the routediscovery process. DP is shown to reduce the number ofbroadcast transmissions when compared to standard DP.Because DP requires the two-hop neighborhood todetermine the forwarder list, we built a neighbor protocolas part of AODV. By making the neighbor protocol partof AODV, the result is a more accurate view of the localtopology, and therefore more accurate is thedetermination of the forwarder list. We also proposed aAPDP algorithm which is an enhanced version of PDP.Here we have shown the performance comparison ofAODV with DP algorithm. Future scope of the work is tocompare to the performance of AODV with PDPalgorithm and compare the performance of both.REFERENCES[1] Brad Williams and Tracy Camp. Comparison ofbroadcasting techniques for mobile adhoc networks. InMobiHoc ’02: Proceedings of the 3rd ACM internationalsymposiumon mobile ad hoc networking & computing,pages 194–205. ACM Press, 2002.[2] C. Perkins. Ad-hoc on-demand distance vector routing. InProceedings of the IEEE workshop on mobile computingsystems and applications, pages 90–100, 1999.[3] Charles E. Perkins and Elizabeth M.Royer. Ad-hoc ondemanddistance vector routing. Technical report, SunMicro Systems Laboratories, Advacnced DevelopmenetGroup, USA.[4] David B Johnson and David A Maltz. Dynamic sourcerouting in ad hoc wireless networks.In Mobile Computing,volume 353, pages 153–179. Kluwer AcademicPublishers,1996.[5] Dipti Joshi, Sridhar Radhakrishnan, and ChandrasekheranNarayanan. A fast algorithm for generalized networklocation problems. In Proceedings of the ACM/SIGAPPsymposium on applied computing, pages 701–8, 1993[6] H. Lim and C. Kim. Flooding in wireless ad hoc networks.Computer Communications, 24(3–4):353–363, February2001[7] Jie Wu and Fei Dai. Broadcasting in ad hoc networks basedon self-pruning. In IEEE INFOCOM, pages 2240–2250,2003.[8] Jorge Nuevo, “A comprensible glomosim tutorial”,March4,2004http://appsrv.cse.cuhk.edu.hk\~hndai\research.[9] Katia Obraczka, Kumar Viswanath, and Gene Tsudik.Flooding for reliable multicast in multi-hop ad hocnetworks. Wireless Networks, 7(6):627–634, 2001.[10] Marco A. Spohn and J. J. Garcia-Luna-Aceves. Enhanceddominant pruning applied to the route discovery process ofon-demand routing protocols. In Proceedings of the 12thIEEE international conference on computercommunications and networks, pages 497–502, October2003[11] Mobile Ad Hoc Networking:Routing Protocol PerformanceIssues and Evaluation Considerations Request ForComments 2501 available at http://www.ietf.org[12] Ning.P. and Sun.K. How to misuse aodv: a case study ofinsider attacks against mobile ad-hoc routing protocols.Technical report, Comput. Sci. Dept., North Carolina StateUniv., Raleigh, NC, USA, 2003.[13] Peng-Jun Wan, Khaled M. Alzoubi, and Ophir Frieder.Distributed construction of connected dominating set inwireless ad hoc networks. In IEEE INFOCOM, pages1597–1604, June 2002.[14] Rajive Bagrodia. README GloMoSim Software.University of California, Los Angeles - Department ofComputer Science. Box 951596, 3532 Boelter Hall, LosAngeles-CA 90095-1596 / rajive@cs.ucla.edu.[15] Teresa W. Haynes, Stephen T. Hedetniemi, and Peter J.Slater, editors. Fundamentals of Domination in Graphs.Marcel Dekker, Inc., 1998.[16] Xiaoyan Hong, Kaixin Xu, and Mario Gerla. Scalablerouting protocols for mobile ad hoc networks. 2002.[17] Yu-Chee Tseng, Sze-Yao Ni, Yuh-Shyan Chen, and Jang-Ping Sheu. The broadcast storm problem in a mobile ad hocnetwork. Wireless Networks, 8:153–167, 2002. H. Lim andC. Kim. Flooding in wireless ad hoc networks. ComputerCommunications,24(3–4):353–363, February 2001-04-15).[18] Z. Haas. A new routing protocol for the reconfigurablewireless network. In Proceedingsof the IEEE internationalconference on universal personal communications, pages562–566, October 1997.[19] A Nagaraju , Dr S.Ramachandram “Adaptive PartialDominating Set Algorithm For Mobile Ad-Hoc Networks”, Presented in the international conference ACM,COMPUTE-09, Jan 9 th -11 th ,2009,Banglore.[20] J.Hass and M.R.Pearlman, “The Zone RoutingProtoco(ZRP) for Ad Hoc Networks,” draft-ietf-manetzone-zrp-02.txt,workin progress,June 1999[21] T.Clausen, P.Jacquet,A.Laouiti, P.Muhlethaler,A. Q andL.Viennot. Optimized link state routing protocol for adhocnetworks. In Proceedings of the IEEE internat multitopics conferences. 2001© 2010 ACADEMY PUBLISHER

86 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Bee-Inspired Routing Protocols for MobileAd HOC Network (MANET)Deepika ChaudharyChitkara Institute of Engg & Technology JhanslaTeh. Rajpura,Distt Patiala-140401,Punjabdeepika.chaudhary@chitkara.edu.inAbstract – Mobile AdHoc networks( MANETs) are receivinga significant amount of attention from researchers. Thispaper provides a high light on a new and very energyefficient algorithm for routing in Manets BeeAdHoc. Thisalgorithm is a reactive source routing algorithm andconsumes less energy as compared to other existing state ofart routing algorithms because a fewer control packets forrouting are sent as compared to other networks.I. INTRODUCTIONMobile Ad HOC networks (MANET’S) are networks inwhich all nodes are mobile and communicate with each other viawireless connections. Nodes can join or leave the network at anytime. There is no fixed infrastructure. All nodes are equal andthere is no centralized control or overview. There no designatedrouters all nodes can serve as routers for each other and datapackets are forwarded from node to node in multihop fashion.Since a few years researcher’s interest in MANETS havebeen growing and especially the design of MANET routingprotocols has received a lot of attention. One of the reasons isthat routing in MANETS is particularly challenging task due tothe fact that the topology of the network changes constantly andpaths, which were initially efficient, can quickly becomeinefficient or even infeasible. Moreover control information flowin the network is very restricted. This is because the bandwidthof the wireless medium is very limited and the medium isshared: nodes can only send or receive data if no other node issending in their radio neighborhood. It is therefore important todesign algorithm that are adaptive robust & self-healing. Natureself-organizing systems like insect societies show precisely thesedesirable properties. Making use of a number of relativelysimple biological agents like ants, a variety of differentorganized behavior are generated at the system level from thelocal interaction among the agents and with the environment.The robustness and effectiveness of such collective behaviorswith respect to variations of environment conditions are keyaspects of their biological success. This kind of systems areoften referred to with the term swarm intelligence. Swarmsystems have recently become a source of inspiration for designof distributed & adaptive algorithms.II. DESIGN ISSUES OF ROUTING ALGORITHMSThe most important challenge in designing algorithms forMANETs are mobility and limited battery capacity of nodes.Mobility of nodes results in continuously evolving newtopologies and consequently the routing algorithms have todiscover or update the routes in real time but with small controloverhead . The limited battery capacity requires that the packetsif possible be distributed on multiple paths , which would resultin the depletion of batteries of different nodes at an equal rateand hence as a result the life time of networks would increase[1].Therefore an important challenge in Manets is to design arouting algorithm that is not only energy effieicent but alsodelivers performance same or better than existing state of artrouting protocols.III. CLASSIFICATION OF ROUTING ALGORITHMS INMANETSThe routing algorithms for Manets can be broadly classifiedas proactive algorithms or reactive algorithms. Proactivealgorithms periodically launch control packets which collect thenew network state and update the routing tables accordingly. Onthe other hand, reactive algorithms find routes on demand only.Reactive algorithms looks more promising from the prespectiveof energy consumption in Manets. Each category of abovementioned algorithms can further be classified into hostintelligent or router intelligent algorithms . A few reactivealgorithms are DSR ( Dynamic Source Routing ) which is hostintelligent algorithm while AODV( AdHoc On DemandDistance Vector Routing) which is a router –intelligentalgorithm.However these algorithms are not designed for energyefficient routing. Here in this study an deep insight is providedon BeeAdHoc which delivers performance same as better thanthat of DSR,AODV but consumes less energy as compared tothem. The algorithm achieves these objectives by transmittingfewer control packets and by distributing data packets onmultiple paths.IV. EXISTING WORK ON NATURE –INSPIRED MANETROUTING PROTOCOLSThe first algorithm which presents a detailed scheme forMANET routing based on ant colony principles is ARA [3]. Thealgorithm has its roots in ABC AND AntNet Routing algorithmsfor fixed networks and are inpired by the pheromone layingbehavior of ant colonies . AntHocNet, which is hybrid algorithmhaving both reactive and proactive components have also beenproposed.This algorithm tries to keep most of features of theoriginal AntNet and shows promising results in simulationcomared to AODV Termite is another MANET routingalgorithm inspired by termite behaviour .Here no special agentsare needed for updating the routing tables rather data packets aredelegated this task.V. OVERVIEW OF BEEADHOC ARCHITECTUREBeeAdHoc is an on-demand multi path routing algorithm formobile adhoc networks inspired from the foraging principles ofhoney bees[4]. BeeAdHoc works with types of agents: packers,scouts foragers and swarms. The packers locate a forager andhand over the data packet to the discovered forager. Scoutsdiscover new routes from the launching node to the destination© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.86-88

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 87node through broadcasting principle and an expanding time tolive (TTL) timer. Foragers, the main workers.The architecture of the hive is shown in Fig 1 where theentrance floor is an interface to the lower MAC layer, while thepacking floor is an interface to the upper transport layer. Thedance floor contains the foragers and the routing information toroute locally generated packets. The functional characteristics ofeach floor composing the hive are explained in the following.floor from the MAC layer. If the packet is the forager and th ecurrent node is its destination, then the forager is forwarded tothe packing floor.; otherwise, it is directly routed to the MACinterface of the next hop node. If the packet is a scout, it isbroadcast to the neighbor nodes if its TTL timer has not expiredyet or if the current node is not its destination. The informationabout the ID of the scout and its source node is stored in a locallist. If a replica of previously received scout arrives at theentrance floor , it is removed from the system. If a forager withthe same destination as the scout already exists on the dancefloor, then the forager’s route to the destination is given to thescout by appending it to the route held so far by the scout.Fig 1. Overview of BeeAdHoc ArchitectureVI. PACKING FLOORThe packing floor is an interface to the upper transport layer(e.g, TCP or UDP) . Once a data packet arrives from transportlayer, a matching forager for it is looked up on the danced floor .If a forager is found then the data packet is encapsulated in itspayload . Otherwise, the data packet is temporary bufferedwaiting for a returning forager. If no forager comes back withina certain predefined time, a scout is launched which isresponsible for discovering new routes to the packet destination .Figure 2 explains the series of action performed at packing floor.Fig 3: The EntranceVIII. DANCE FLOORFig 2 : The Packing FloorVII. ENTRANCEThe function performed in the entrance are shown in figure.3. The entrance is an interface to the lower level MAC layer.The entrance handles all incoming and out going packet. Actionon the dance floor depends on the type of packet entered theThe dance floor is the heart of the hive because it maintainsthe routing information in the form of foragers. The dance flooris populated with routing information by means of mechanismreminiscent of the waggle dance recruitment in natural bee hivesonce a forager returns after its journey, it recruits new forager by“dancing” according to the quality of the path it traversed.A lifetime forager evaluates the quality of its route based onthe average remaining battery capacity of the nodes along itsroute. The central activity of the dance floor module consists ofsending a matching forager to the packing floor in response to arequest from a packer. The foragers whose lifetime has expiredare not considered for matching. If multiple path be identifiedfor matching, then a forager is selected in a random way. Thishelps in distributing the packets over multiple paths, which inturn serves two purpose : avoiding congestion under high loadsand depleting batteries of different nodes at comparable rate. Aclone of the selected forager is sent to the packing floor and thenumber of permitted clones. If the dance number is 0 then theoriginal forager is sent to the packing floor removing it in thisway from the dance floor. This strategy aims at favoring youngover old foragers.If the last forager for a destination leaves ahive then the hive does not have any more route to thedestination. Nevertheless if a route to the desination still existsthen soon a forager will be returning to the hive if no foragercomes back within reasonable amount of time, then the node has© 2010 ACADEMY PUBLISHER

88 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010probably lost its connection to the destination node. In this wayfewer control packets are transmitted resulting in less energyexpenditure.packets over different routes rather then always sending them onthe best routes.X. FUTURE ASPECTSDesign and development of routing protocols for Mobile AdHoc Networks (MANETs) is an active area of research. Thestandard practice among researchers working in this emergingdomain is to evaluate the performance of their routing protocolsin a network simulator. In future a mathematical models of twokey performance metrics, routing overhead and route optimality,of BeeAdHoc MANET routing protocol using a simulator can bestudied . One of the key components of our BeeAdHoc model isthe collision model at Medium Access Control (MAC) layer. Todesign the mathematical expressions of the performance metricscan also provide insight about the behavior of BeeAdHoc inparticular, and a typical ad hoc routing protocol in general.REFERENCESIX. CONCLUSIONThe simplicity of BeeAdHoc which results from its simplerarchitecture and its using a smaller number of control packetspay off once we look at the energy consumption in transpotingthe packets from their source to their destination. BeeAdHocnetwork employs a simple bee behavior to monitor the validityof the routes.When compared to other algorithms the batterylevel of BeeAdHoc is better because it tries to spread the data[1] Formal Modeling of BeeAdHoc : A Bio-inspired Mobile AdHoc Network Routing Protocol, Muhammad Saleem 1 , SyedAli Khayam 2 and Muddassar Farooq 3.[2] The wisdom of the hive applied to mobile ad-hoc networksWedde, H.R.; Farooq, M.[3] BeeAdHoc: an Efficient, Secure and Scalable RoutingFramework for Mobile AdHoc Networks , Prof. Dr. HorstF. Wedde (wedde@ls3.cs.uni-dortmund.de) ME MuddassarFarooq (muddassar.farooq@ls3.cs.uni-dortmund.de)[4] M. Farooq, “From the Wisdom of the Hive to IntelligentRouting in Telecommunication Networks” PhD Thesis,Dortmund University, 2006.[5] H. F. Wedde and M. Farooq. “The wisdom of the hiveapplied to mobile ad-hoc networks.” In Proceedings of theIEEE Swarm Intelligence Symposium, pages 341–348,2005.[6] P. K. Visscher, “Dance Language”, Encyclopedia ofInsects, Academic Press, 2003© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 89Distance and Frequency based Route StabilityEstimation in Mobile Adhoc NetworksAjay KoulSMVD University/ School of computer Science and Engineering, Katra, IndiaEmail:ajay.kaul@smvdu.ac.inR. B. Patel and V. K. BhatM.M.University/Department of Computer Science and Engineering, Ambala, IndiaSMVD University/School of Applied Physics and Mathematics, Katra, IndiaEmail :{ patel_r_b, vijaykumarbhat2000}@yahoo.comAbstract-This paper discusses the link stability estimationfor Mobile Ad hoc Networks (MANETs). In this approachthe total time for which the link remains connected with theneighboring nodes is estimated. This helps to predict thestability of the route which is required to forward thepackets to the destination. This method does not use theselection of the next hop on the basis of shortest distance,but is based on the time period for which the next hop linkremains connected. The parameters we use, are the distanceand frequency and signal quality. These provide the way fora node to decide the best next hop neighbor and hence aperfect Quality of Service (QoS) is also obtained.The rest of the paper is organized as follows.In Section II we discuss the related works. In Section III,the node distance, position and mobility is estimated.The quality of signal along with the algorithms to beexecuted on intermediate and destination nodes arepresented in Section IV. In Section V the practicalscenario of estimating the parameters like distance,frequency, mobility, RSSI, etc. and their performanceresults are obtained. Finally the article is concluded inSections VI.Index Terms— Route stability, frequency, distance,MANETs, QoS.I. INTRODUCTIONAn ad hoc network is a dynamic multihop wirelessnetwork that is established by a set of mobilenodes. Such networks are, therefore, suitable for theenvironments where it is a difficult to create a fixedinfrastructure. In this network, mobile nodes randomlymove and communicate over radio channels. If twomobile nodes are in a radio transmission range, they cancommunicate with each other directly, otherwise, thesource node sends/receives the packets via someintermediate nodes. Hence a proper routing algorithm isrequired to route the packets from source to destination.Most routing algorithms, like in [1] estimate the linkstability based on the distance between the twoneighboring nodes. In [2] the electric field as a parameteris used to find the stable route to the destination. Thealgorithm given in [3], finds the best route to thedestination based on link perdurability. These, however,do not lead to the reliable solution in Mobile Ad hocNetworks (MANETs), as the environments are dynamicand hence the route estimation depends on the factors likemobility and direction as well. In this paper we havetaken these factors into account and the next hop route isselected not on the basis of distance only but on mobilityas well. The quality of strength is also taken intoconsideration. Our results show that our method can alsobe applied in the selection of the best neighboring node.II. RELATED WORKMany routing protocols in the past have been proposed onroute stability. In [4] the technique of signal stability isused for adaptive routing in Mobile Ad hoc networks. Inthis approach the on demand longer lived routes arediscovered based on signal strength and location stability.The signal strength of the neighboring nodes are detectedby sending beacons, and based on that, the classificationof strong and weaker channels are established. This isfurther strengthened by choosing a channel which hasexisted for a longer period of time. This signal strengthfeature in combination with the location stability helps inselection of the next hop neighbor and hence the overallroute is established. The protocol mentioned even thoughgives good results in establishing stable route, however,fails when the node density increases or the node mobilityincreases. The Global positioning system based reliableroute discovery proposed in [5] used a different approach.The algorithm discovers routes based on two zones, thestable zone and the caution zone. The zones are decidedbased on the location and the mobility of the nodes usingthe Global positioning system (GPS). The stable zone andthe caution zone change dynamically depending on themobile nodes speed and direction information. Themobile nodes speed, direction and position is estimatedbased on GPS system. This method is the perfect methodof determining the stability of the route as the criticalparameters like direction and mobility is determined. Themethod mentioned however, has a serious drawback ofadditional cost involvement and more power requirementfor the GPS based device. In [6] a model to find the best© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.89-95

90 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010route based on link lifetime has been proposed. In thismodel the edge effect has been explored to find morestable route to the destination. The lifetime and thestability of the route is calculated by eliminating the edgeeffect to reduce route maintenance and route overheads.The disadvantage of this model however is that eachmethod proposed under the model requires either pilotsignal generation and monitoring of the pilot signal of theother nodes or monitoring of signal strength of the othernodes. With these weaknesses, the stable route still getsestablished and performance of the network also getsincreased. Lifetime Prediction Routing (LPR) proposed in[7] also was suggested to find the stability of the route. Inthis the service life of the MANET is predicted andcorrespondingly the route is established. The servicelifetime is predicted based on the past activity of thebattery lifetime of the node. A simple moving averagepredictor is used to keep track of last N values of residualenergies and the corresponding time instances for the lastN packets received/ relayed by each mobile node. Thisway LPR not only captures the remaining batterycapacity but also accounts for rate of energy discharge.This way the route stability and lifetime is estimatedbased on the battery life which is also an important factor.This method has some serious concerns like, when thenode mobility increases it becomes difficult to predict thelifetime of the node and also the use of LPR involvescertain overheads. One more model to predict the linkstability was proposed in [8]. This model was named asthe signal stability based routing protocol (SSA + ) model.SSA + is the enhanced version of SSA for finding theroute stability in MANETs. In this method a route ismaintained with the help of active neighbouring nodes.The neighbours are considered active if it relays ororiginates at least one packet within the most recentactive timeout period. The SSA + solution was mainlyoffered to remove the problems of high node density andhigh mobility and the low node density and low mobility.In high node density and high mobility scenario themobility is high, the probability of link failures remainalso high. To cater to the problem the signal strength ofthe link is estimated and classified into the categories ofweak, normal and strong signals. The nodes exchangeinformation regarding the signal strength by sending thehello packets and based upon the signal strength valuestored in the link state table, the life time of the route isdetermined. In low node density and low mobility thesignals of the nodes remain weaker and the stability andthe route lifetime is maintained based on two importantconditions.1) The distance between any two nodes getsshorter for the past few clicks.2) Secondly the distance between any two nodesgets longer but the distance changes largerslowly for the past few clicks (the condition oflow mobility).The method above mentioned above, even thoughshowed better results compared to SSA in terms of lownode density and low mobility and high node density andhigh mobility has the drawback of predicting the lifetimeof the route in a complicated way as the node has tomaintain the Link stability table which adds lots ofoverheads in terms of route control and maintenance andwhich are difficult to maintain. In [9] [10] and [11],algorithms are proposed as DV-Hop, Hop-TERRAIN,and link stability with dynamic delay prediction, todetermine the location of nodes based on hop counts andthe appropriate route to the destination. The hop countsprovided an estimate for the overall distance between thenodes and the dynamic delay ensured stability. Thesehowever, were mainly focusing on the Quality of service(QoS) based route establishment and hence were silent onthe route lifetime parameters like mobility and batterypower. A Novel route metric based on the fragility of theroute was also proposed in [12]. In this approach thedynamic nature of the route is captured by studying thedistance variation of the next hop neighbor in terms ofexpansion and contraction. A distributed algorithm isexecuted in every node to calculate the relative speedestimate of the neighbours. The destination node gets theinformation of every route and selects the best routebased on the route fragility coefficient (RFC) which inturn depends on the cumulative expansion metrics (CEM)and cumulative contraction metrics (CCM). Thisapproach finds the stability of the route and is good interms of not requiring time bound measurements, like theglobal positioning system. This method however, fails topredict the node mobility and includes the destinationlatency. The estimation of link quality and residual timein vehicular Adhoc networks proposed in [13] is also themethod to determine stability of the route in MANETs bypredicting the active time of the link. In this method thesignal processing along with empirical decompositionand robust regression is used to predict the link qualityand the residual time. This method unlike the methodproposed in this paper uses a three stage approach whichinvolves lots of complexity in terms of practicaldeployment. This also involves additional burden ofcalculating the various essential parameters mentionedunder signal processing, empirical decomposition androbust regression methods. The above mentionedliterature even though provides excellent way ofestimating the distance and the location of neighboringnodes through different techniques, however, are mostlysilent on the mobility issue. This paper provides the routestability estimation based upon the neighboring nodedistance ,mobility and signal strength.III. DESCRIPTIONMobile Ad hoc Networks (MANETs) being dynamic innature create challenges in terms of Quality of Service(QoS) at each level from application to physical. Inphysical layer level the mechanism to predict the lifetimeof the neighboring node increases reliability from sourceto destination delivery. To achieve this we have identifiedthe following three parameters for next hop path selection1) Node Distance and position2) Mobility3) Signal quality© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 91A. Node Distance and positionLet two circles with radii R 1 and R 2 intersect in a regionas shown in Fig. 1 Let the circles be centered At A (0, 0)and B(d, 0). Then equations of the circles are( d )2 2 2x + y = R1(1)2 2 2x − + y = R2(2)This implies that2 2 2 2 2x − 2dx+ d − x = R2− R1(3)⎛ 2 2 2 ⎞⎜ d − R + Rx =2 1 ⎟⎜ 2d⎟⎝⎠Figure 1. Showing the wireless range intersection.(4)Now y is half of the length of chord connecting the twocircles, we have2 2y = R1− x2 2 2 2 2 224dR1− ( d − R2+ R1)y = (5)24dTo find out d 1 and d 2 the distances from centre on nodesto the circle intersection point, we use the relation22 2 2( d − R R )d2+11= (6)2dd22 2 2( d + R2− R1)= (7)2dTo find the distance d, and position parameters x, ywe assume the following:1) That all the nodes are of equal strength andtechnical specifications2) That the radii are known and are same for allthe nodes because of same transmission power.With these assumptions and from (1-7), the distance canbe found using the method as given in [14], known asround trip time based method. This works on the conceptthat every data packet can be acknowledged. Under thiswork the time span from the moment at which a packetstarts to occupy the wireless medium to the time at whichthe immediate acknowledgment is received is measuredand denoted by T R . The time duration between thereception of a data packet and issuing the correspondingimmediate acknowledgment is also measured anddenoted by T L . The distance is computed based on therelation between the distance traveled and the speed oflight as follows:( TR−TL)d = × C2(8)Where C= 3 × 10 8 is the speed of light. Now solvingfor d1, d2, x,y from (7) and (8), the positionP ( α , β ) of the neighboring node can be estimated. Nowlink stability L s is, given by:⎧1,if d < µ and | AP | ≤|AC |L s = ⎨⎩0,if d < µ and | AP | ≤|AC |WhereAP =2 2α + β2 2AC = m + nand µ and C(m, n) are respectively the maximumpermissible distance and maximum co-ordinate positionallowed to communicate between the nodesB. MobilityMeasurement of distance and position as mentionedabove are easy to compute and provides the route stabilityespecially when the nodes are static. In case the nodes aredynamic the mobility plays an important factor. Thedistance and position in mobile environment will providethe feasibility of communication but will not provide thelife time of the route with the neighboring node. To findmobility, we calculate the frequency of the neighboringnode from the following equations using [10].⎡ v ⎤f r = f e ⎢ ⎥(9)⎣v+ v sr ⎦Where in (9), f r is the received frequency, f e is theemitted frequency, v is the speed of the waves in themedium and v sr is the radial component of the velocity ofthe neighboring node with respect to the medium(positive if moving away from the observer, negative ifmoving towards the observer). A similar analysis for amoving observer and a stationary source yields theobserved frequency from the following equation (thereceiver's velocity being represented as v r ):⎡ v ⎤f r = f e ⎢ ⎥(10)⎣ v + v r ⎦where the same convention applies. We note that v r ispositive if the observer is moving away from the sourceand negative if the observer is moving towards the© 2010 ACADEMY PUBLISHER

92 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010source. These can be generalized into a single equationwith both the source and receiver moving as given below:⎡ v ⎤f = ⎢ − srr f e 1 ⎥(11)⎣ v + v sr ⎦Where v sr is the source to receiver velocity radialcomponent. Now since the source nodes have equalpower so frequency of transmission will remain same forall the nodes within the vicinity. From (11) above we getthe velocity of the neighboring nodes as⎡ f − f ⎤v = e rsr v⎢⎥(12)⎣ f r ⎦Now overall time period for which the link remainsestablished will depend on the following relationdT = (13)v srwhere T in (13) is defined as the time period for whichthe link with the neighboring node will remainestablished. The time period depends on d because if thenode covers the maximum distance away from the othernode, then a small velocity of it in the opposite directionwill disconnect the link. The time T will be more if it isnegative as the velocity will be having +ve or –ve signdepending upon the direction of motion.C. Signal QualityThe distance, and position plays an important role infinding mobility and hence the link stability. However,the stability of the link also depends on the quality ofsignal as well. In Mobile Adhoc Networks (MANETs)the signal quality plays an important role in selecting theroute to the destination. The parameters like distance andmobility proposed above effects the signal quality,however in spite of the feasible distance and mobilitythreshold values, the node may not receive from theneighboring node the good quality signal due to noisysurroundings. When there is a signal transmission from anode with certain power in the noisy environment, the Biterror rate of the signal increases apart from the normalpath loss in the medium. The receiver on the basis of thereceiver sensitivity, and the threshold Signal to NoiseRatio (SNR), predicts the quality of signal. Overall it isthe Receiver signal strength indicator (RSSI) whichprovides the details of the quality of signal received.RSSI which is basically a measurement of how well theradio is receiving or 'hearing' data to determine thequality of signal received. It's typically measured in -dBm, which is the power ratio in decibel (dB) of themeasured power referenced to one milliwatt (mW).Normally in the real testing environments the RSSI above-60 dB is considered the threshold required to performgood networking functions and any value higher than thatis the stable value. We denote this value by .Therefore overall link stability which is denoted by L SSOshould satisfy L S , T and M RSSI the three importantparameters to determine the stability of the route.IV. ROUTINGThe above parameters calculated can be used to enhancethe routing procedures already available. To incoorporatethe changes, the nodes needs to be modified.Below are the node level modifications suggested whenthe node behaves as, the intermediate and the destinationnode.A. Intermediate node operationWhen the intermediate node receives the RREQ packetfrom the source or from any of its neighbors it providesthe additional information of the distance and thefrequency in the RREQ packet and forwards it to thedestination. The Algorithm 1 provides the operationaldetails to be performed at each intermediate node.Algorithm 1 Algorithm to be executed in the node1 Set , (m, n), M RSSISet k stable link time threshold2 Measure d3 if d < µ L s = 1 Goto xElse d > µ L s = 0 stop.3 x: calculate RSSI4 if RSSI < M RSSI goto 8 else5 calculate v sr5 if vsr -ve calculate TElse if +ve calculate T6 if T < k reject6 Else if T > k select and modify RREQ7 stopB. Destination node operationThe destination node on receiving RREQ packets fromthe several nodes sends the reply to the source byselecting the best path on the basis of the T, M RSSI value.Algorithm 2 provides the process details of thedestination node when RREQ arrives from differentnodes of the destination node.Algorithm 2 Process execution when RREQ arrives1 Analyze RREQ2 Extract, d, v sr M RSSI*3 Compared with d* and M RSSI with M RSSI * and v sr with v sr*4 if d > d* and M RSSI < M RSSI*and v sr > v sr*Reject RouteElse select RouteThe source node when sending the RREQ packet,calculates before, the distance, mobility and SNR of theneighboring nodes. It then embeds the information in theRREQ packet and sends it to the destination via. Nexthop neighbor.© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 93V. RESULTSThe experiments were performed in the indoorenvironments. In the experiment, we used one wirelessaccess point namely Wireless-G WAP54Gv.2 of Linksysmake and a WLan laptop. The access point and theLaptop were supporting the wireless 802.11 b/g modelprotocols. For sending the ICMP packets, we used aGraphic tool namely Multiping Grapher to send the pingrequests every second.The payloads in bytes were keptconstant at 512bytes. The purpose of doing so was to seethe graphic outcome of the response. Figures 2 and 3show the details of the ping at a max of 6ms and 7ms byvarying the indoor distance from maximum to minimumand vice versa.TABLE I.Velocity and Frequency Recorded ReadingsS.No velocity in m/s Frequency in GHz1 -2 2.41402 -3 2.42123 -4 2.42834 -5 2.43565 -6 2.44216 -7 2.44977 -8 2.45738 -9 2.46439 -10 2.471910 -11 2.479311 -12 2.4868Figure. 2 Ping Response at 6ms MaxThe physical and the operational mode rates have beenfixed at 802.11g and 54mbps, 11mbps. The distance d ascalculated in (8) is measured from the tool Snuffle.From this tool the and is computed from theMAC time stamps recorded on the transmitted requestand received acknowledgment packets and by subtractingthe MAC time stamps of the received reply and the issuedacknowledgment packets.Figure. 3 Ping Response at 7 ms Max.Figure 4. shows the distance calculated vs the differencein MAC delay. The calculated distance is however not theactual physical distance. The intension here is to checkthe distance between the nodes and to compare it with thethreshold maximum allowable radial distance in this casewe have fixed as . Now that the distance between thenodes is calculated, the mobility can be calculated byspecifying the frequency of the source at rest which is 2.4GHz and wave velocity 300,000 Kms/second. Theexperiments were carried in the indoors were the speed ofthe access point were varied and the frequency it wascalculated by using the above equations and theDoppler’s effect calculator.Frequency in GHz2.52.482.462.442.422.42.382.361 2 3 4 5 6 7 8 9 10 11Speed in m/sFigure. 5 Frequency vs SpeedThe data collected are as shown in Figure 5. Thevelocities in the table 1 above are shown negative thisindicates that the node is approaching the other nodewhich is recording the frequency. By calculating thevelocity we can find the link stability as per the equationsgiven above in (13) and (14).A. Received Signal Strength IndicatorTo calculate the link stability based on RSSI parameter,we used a graphical tool called Wirelessmon. This toolprovided the details of the RSSI Value recorded and thesignal strength percentage which is proportional to theRSSI value. The more the percentage of signal strengththe more the RSSI value closer to zero. Some of theGraphs plotted are shown from Fig. 6 to Fig. 8. Thegraphs clearly show that the percentage of signal receivedis good when the distance is only 1m and hence the RSSIvalue approaching –ve value towards zero. As soon as thenode is moving away from the access point and somealuminum wall partition included in between, It showsthe decline in the signal and hence in the RSSI value aswell.Delay in µ SecSignal strength (%)Distance in Meters x10Figure. 4 Calculated distance vs DelayTime (Seconds)Figure. 6 Signal strength % vs time in seconds© 2010 ACADEMY PUBLISHER

94 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Signal strength (%)strength (%)SignalTime (seconds)Figure. 7 Signal strength % vs Time in seconds distance=3m and one aluminium partition in between the nodesTime ( seconds )Figure. 8 Signal Strength % vs. time in sec with distance = 6m andtwo aluminum partitions between the nodes.VI. CONCLUSIONSWe have presented an approach on how to measure thedistance and position based on the propagation time ofIEEE 802.11 packets. We used commercial WLAN cardof linksys make supporting IEEE 802.11b and 802.11g.The distance and the position measure however is logicaland has no relation with the physical distance or positionin the co-ordinate system. The frequency measurementresults reveal the mobility of the neighboring nodes. Thedistance parameter can be used to detect the location ofthe node especially when two nodes are static and istherefore, sufficient to determine next hop link stability.The frequency measurement to determine the logicalmobility is however important to determine the next hoproute stability if any one or both of the nodes are mobile.The signal quality also adds the QoS factor. This paper inspite of exploring three important factors like distanceand mobility RSSI value however is silent on the otherfactors like S/N ratio, BER and overall signal quality. Thedistance factor can somehow presume the strength ofsignal but cannot provide the details of the qualityparameters like S/N ratio, BER. These parameters S/Nratio and BER combined with the above mentionedparameters like distance and frequency may be used tofind the best route life in near future. The algorithmsdepicted also will be incorporated in the nodes and arouting protocol will be used to see the performanceenhancement of this method over others.REFERENCES[1] Nam T. Nguyen, Ady Wang, Peter Reiher, GeoffKuenning “ Electric Field Based Routing : AReliable Framework for Routing in MANETs ”,ACM Mobile Computing and CommunicationsReview, USA, 8(2): 35-49, 2004.[2] Sun Xuebin, Zhou Zheng, “ Link Perdurability basedRouting for Mobile Ad hoc Networks ”, Journal ofElectronics, China, 20(4): 299-304, 2004.[3] M. Royer and C.-K. Toh, “A Review of currentRouting Protocols for Ad Hoc Mobile WirelessNetworks”, IEEE Personal Communication Magazine,6(2): 46-55, 1999.[4] Rohit Dube, Cynthia D. Rais, Kuang-Yeh Wang, andSatish K. Tripathi, “Signal Stability-Based AdaptiveRouting (SSA) for Ad Hoc Mobile Networks,” IEEEPersonal Communications, 4(1): 36-45 1997.[5] Young Joo Soh, Won Kim, Dong Hee kwon, “GPS basedReliable Routing algorithms for Adhoc networks” inProceedings of IEEE International Conference onCommunications, Helsinki, Finland, June 11- 14, 2001,pp.3191-3195[6] G. Lim, K. Shin, S. Lee, H. Yoon and J. S. Ma,“Link Stability and route lifetime in ad hoc wirelessnetworks”, in Proceedings of International Conferenceon Parallel Processing workshop (ICPPW’02),Vancouver, BC, Canada, August 18-21, 2002, pp. 116-123.[7] M. Maleki, K. Dantu, and M. Pedram, "Lifetimeprediction routing in mobile ad-hoc networks," IEEEWireless Communications and Networking Conference,USA, March 16-20, 2003, vol.2, pp.1185–1190.[8] W.-F. Wang and P.-H. Shih, "Study on an enhanced linkstabilitybased routing scheme for mobile ad hocnetworks," in Proceedings of 3rd Annual IEEECommunication Society Conference on Sensor and Ad HocCommunications and Networks (SECON'06), Reston, VA,USA, vol.3, September 25-28, 2006, pp. 797–802.[9] F. Erbas, J. E.Garcia, K. Jobmann, “Position- based QoSrouting in mobile ad hoc networks problem statementand a novel approach”, in Proceedings of 3 rd IEEEInternational Conference on performance, Computingand communications, Pheonix, AZ, USA, April 15- 17,2004, pp. 619- 623.[10] S.R. Medidi, K.H. Vik, “Quality of service- Awaresource-Initiated ad-hoc routing ”, in Proceedings ofIEEE SECON Conference Pullman, USA, October 4-7,2004, pp. 108-117.[11] Peng Yang, Biao Huang, "QoS Routing Protocol Based onLink Stability with Dynamic Delay Prediction inMANET”, IEEE Pacific-Asia Workshop onComputational Intelligence and Industrial Application(PACII), Wuhan, China, vol. 1, December 19-20, 2008,pp.515-518..[12] O. Tickoo, S. Raghunath, Kalyanaraman S, “Routefragility: a novel metric for route selection in mobile adhoc networks”, in Proccedings of the 11 th IEEEinternational conference on networks (ICON), SydneyNSW, Austrailia, september 28- 1 st october 2003, pp. 537-542.[13] Nikoletta Sofra, K. K. Leung, “ Estimation of Link qualityand Residual Time in VANETs”, in Proceedings of IEEEwireless communications and Networking Conference(WCNC), Sydney, Austrailia, 31 March- 3 rd April, 2008,pp.2444-2449.[14] Murad Abusubaih, Berthold Rathke, and AdamWolisz, “A Dual Distance Measurement scheme forIndoor IEEE Wireless Local Area Networks”, inProceedings of the 9 th IFIP/IEEE InternationalConference on Mobile and wireless CommunicationNetworks (MWCN’07), Ireland, Cork, September 9-21,2007, pp. 121-125.© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 95[15] Branislav Kusy, Akos Ledeczi, Xenofon,Koutsoukos, “Tracking Mobile nodes using RF Dopplershifts”, in Proceedings of the 5 th InternationalConference on Embedded networked sensor systems,Sydney, NSW, Australia, November 6-9, 2007, pp.29-42.Ajay koul received his B.E. degree from BangaloreUniversity, Bangalore, Karnataka, India in Electrical& Electronics in 1997 and MTech from HyderabadUniversity, Hyderabad, Andhra Pradesh, India inComputer Science in the year 2001.He is currently a Ph.D. student in the School ofComputer Science at SMVD University Katra, J&K,India. His research interests include wirelessnetworks, Image processing and data storage.Dr. R. B. Patel received his PhD from IIT Roorkee,Roorkee, Uttarakhand, India, in Computer Science &Engineering. PDF from Highest Institute ofEducation, Science & Technology (HIEST), Athens,Greece, MS (Software Systems) from BITS, Pilani,RajasthanHe is in teaching and Research & Development since 1991. Hehas supervised several M. Tech, M. Phil and PhD . He haspublished more than 95 research papers inInternational/National Journals and Refereed InternationalConferences. His current research interests are in Mobile &Distributed Computing, Mobile Agent Security and FaultTolerance, development infrastructure for mobile & Peer-To-Peer computing, Device and Computation Management,Cluster Computing, etc. He has written numbers books forengineering courses (These are “Fundamentals of Computingand Programming in C”, “Theory of Automata and FormalLanguages”, “Expert Data Structures with C,” out of these someare recommended by Indian Society for Technical Education(ISTE), India.Dr. Patel received several research awards, like best researchpaper award at BIS 2003 Colorado, Spring, USA and is amember of various International Technical Societies such asIEEE-USA, Elsevier-USA, Technology, Knowledge & Society-Australia, WSEAS, Athens.Dr. V.K. Bhat received his MscDegree from KashmirUniversity, Kashmir and Phd Degrees in Mathematics fromJammu University, Jammu, J&K, India.He is currently serving in the School of Mathematics at SMVDuniversity Katra India. He has supervised many researchstudents in the area of Ring theory, Load balancing anddistributed networks. He has published nearly 65 researchpapers mostly in international journals and conferences ofrepute. His current research interests include Graph theory andRing theory.© 2010 ACADEMY PUBLISHER

96 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Investigation of Blackhole Attack on AODV inMANETAnu BalaUniversity Institute of Engg. & Tech Deptt, Panjab University, Chandigarh,IndiaEmail: anubala22@gmail.comRaj Kumari and Jagpreet SinghUniversity Institute of Engg. & Tech Deptt, Panjab University, Chandigarh,IndiaGuru Teg Bahadur Khalsa Institute of Engineering and Technology, Malout, IndiaEmail:rajkumari_bhatia5@yahoo.com and drjagz@gmail.comAbstract—Mobile Ad Hoc Network (MANET) consists of acollection of wireless mobile hosts without the requiredintervention of any existing infrastructure or centralizedaccess point such as base station. The dynamic topology ofMANET allows nodes to join and leave the network at anypoint of time. Wireless MANET is particularly vulnerabledue to its fundamental characteristics such as open medium,dynamic topology, distributed cooperation and constrainedcapability. In this paper we simulate the blackhole attackwhich is one of the possible attacks on AODV routingprotocol in mobile ad hoc networks by the help of networksimulator (NS-2). The simulation results show the packetloss, throughput, and end-to-end delay with blackhole andwithout blackhole on AODV in MANET. We analyzed thatthe packet loss increases in the network with a blackholenode. We also observed that the throughput and end-to-enddelay decreases in the network with a blackhole node.Index Terms—introduction, AODV routing protocol,blackhole attack in AODV, simulation environment andresults, conclusion, acknowledgement, referencesI. INTRODUCTIONWireless networks use some sort of radio frequenciesin air to transmit and receive data instead of using somephysical cables. Wireless networks are formed by routersand hosts. Ad-hoc networks are wireless networks wherenodes communicate with each other using multi-hoplinks. Networks that support mobile wireless ad hocarchitecture are typically called mobile ad hoc networks(MANET). A mobile ad hoc network is formed bymobile hosts. There is no stationary infrastructure or basestation for communication. So the functioning of Ad-hocnetworks is dependent on the trust and co-operationbetween nodes. Ref. [1] Nodes help each other inconveying information about the topology of the networkand share the responsibility of managing the network.Ref. [2] Each mobile node acts as a host whenrequesting/providing information from/to other nodes inthe network, and acts as router when discovering andmaintaining routes for other nodes in the network.Ref. [3] Based on the routing information updatemechanism, routing protocols in ad hoc wirelessnetworks can be classified into three broad categories:Proactive (or table-driven) protocols, Reactive (or ondemand)protocols, and Hybrid routing protocols. Theseare further divided into sub categories. Ref. [3] These arevulnerable to routing attacks. Routing attacks in ad hocwireless networks can also be classified into five broadcategories: Attacks using Impersonation, Modification,Fabrication, Replay, and Denial of Service (DoS). In thispaper, we focus on blackhole attack that belongs tocategory of fabrication attacks.Ref. [4] There are three main routing protocolsproposed for MANET: Ad hoc On-demand DistanceVector (AODV) routing, Dynamic Source Routing(DSDV), and Destination Sequence Distance Vectorrouting protocols. AODV and DSR belong to on-demandrouting protocols and DSDV is a table-driven routingprotocol. These protocols are vulnerable to differentsecurity attacks. In this paper, we use AODV routingprotocol because the AODV protocol is vulnerable to theblackhole attack. So we have simulated the behavior ofblackhole attack on AODV in MANET.II.AODV ROUTING PROTOCOLRef. [5] Ad-Hoc On-Demand Distance Vector(AODV) is a reactive routing protocol in which thenetwork generates routes at the start of communication.Ref. [7] The Ad Hoc On-Demand Distance Vector(AODV) routing protocol described in builds on theDSDV algorithm. AODV is an improvement on DSDVbecause it typically minimizes the number of requiredbroadcasts by creating routes on a demand basis, asopposed to maintaining a complete list of routes as in theDSDV algorithm. Ref. [7] The authors of AODV classifyit as a pure on-demand route acquisition system, sincenodes that are not on a selected path do not maintainrouting information or participate in routing tableexchanges.AODV builds routes using a route request / route replyquery cycle. When a source node desires a route to adestination for which it does not already have a route, itbroadcasts a route request (RREQ) packet across thenetwork. Nodes receiving this packet update theirinformation for the source node and set up backwardspointers to the source node in the route tables. In additionto the source node's IP address, current sequence number,and broadcast ID, the RREQ also contains the mostrecent sequence number for the destination of which the© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.96-100

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 97source node is aware. A node receiving the RREQ maysend a route reply (RREP) if it is either the destination orif it has a route to the destination with correspondingsequence number greater than or equal to that containedin the RREQ. If this is the case, it unicasts a RREP backto the source. Otherwise, it rebroadcasts the RREQ.Nodes keep track of the RREQ's source IP address andbroadcast ID. If they receive a RREQ which they havealready processed, they discard the RREQ and do notforward it. As the RREP propagates back to the source,nodes set up forward pointers to the destination. Once thesource node receives the RREP, it may begin to forwarddata packets to the destination. If the source later receivesa RREP containing a greater sequence number orcontains the same sequence number with a smaller hopcount, it may update its routing information for thatdestination and begin using the better route. As long asthe route remains active, it will continue to bemaintained. If a link break occurs while the route isactive, the node upstream of the break propagates a routeerror (RERR) message to the source node to inform it ofthe now unreachable destination(s). After receiving theRERR, if the source node still desires the route, it canreinitiate route discovery.The rest of this paper is organized as follows. Insection II, we discuss the AODV routing protocol indetail. Section III describes the characteristics of theblackhole attack on AODV. Section IV provides thesimulation environment and results. Finally we concludein section V.III.BLACKHOLE ATTACK in AODVRef. [5] In a blackhole attack, a malicious node canimpersonates a destination node by sending a spoofedroute packet to a source node that initiates a routediscovery. Ref. [2] A blackhole has two properties:1. The node exploits the ad hoc routing protocol,such as AODV, to advertise itself as having a valid routeto a destination, even though the route is spurious, withthe intention of intercepting packets.2. The node consumes the intercepted packets.In an ad hoc network that uses the AODV protocol, ablackhole node absorbs the network traffic and drops allpackets. To explain the blackhole attack we add amalicious node that exhibits blackhole behavior in theFig. 1.node N8 in Fig. 1, and initiates the route discoveryprocess. We assumed node N4 is a malicious node withno fresh enough route to destination node N8. However,node N4 claims that it has the route to the destinationwhenever it receives RREQ packets, and sends theresponse to source node N1. The destination node andany other normal intermediate nodes that have the freshroute to the destination may also give a reply. If the replyfrom a normal node reaches the source node of theRREQ. First, everything works well; but the reply frommalicious node N4 could reach the source node first, ifthe malicious node is nearer to the source node.Moreover, a malicious node does not need to check itsrouting table when sending a false message; its responseis more likely to reach the source node first. This makesthe source node think that the route discovery process iscomplete, ignore all other reply messages, and begin tosend data packets. As a result, all the packets through themalicious node are simply consumed or lost. Themalicious node could be said to form a black hole in thenetwork. In this way the malicious node can easilymisroute a lot of network traffic to itself, and could causean attack to the network with very little efforts on its part.IV.SIMULATION ENVIRONMENT and RESULTIn this section we present a set of simulationexperiments to evaluate the effect of blackhole attack onAODV protocol in MANET. First I have explainedblackhole attack in detail via simulation in NS-2.We havegenerated a small size network with7 nodes in a flat gridof 670m x 670m including blackhole node. We havegenerated a connection between nodes 1 and node 2. Wehave also introduced some movements in our scenario.Duration of the scenario is 60 seconds. Node 1 is thesource node, node 2 is the destination node and node 6 isthe blackhole node. Fig. 2 shows the data flow from node1 to node 2 via intermediate nodes 3 and 4. For someseconds, the link breaks and all data that is send fromnode 1 get lost as shown in Fig.3. Now Fig. 4 shows thatnode 1 again sends the RREQ to all nodes to find route.Nodes further rebroadcast the request if they are not thedestination nodes. Node 6,that is blackhole node, claimsthat it has the route to destination whenever it receivesRREQ packets and sends the response to source node 1.All other nodes that have the fresh route also send areply. But the reply from node 6 reaches the source nodefirst. Node 1 accepts it and ignores all other replymessages and begins to send data packets to node vianode 3, 4, 5 and Node 6 being a blackhole node absorbsall the packets and traffic as shown in Fig. 5.Secondly, to calculate network performance, wesimulate blackhole node behavior in AODV in largenumber of nodes and connections with the help ofNetwork Simulator 2 Ref. [6]. We set the parameters forour simulation as shown in Table 1.Figure 1: Blackhole attack in AODVIn Fig. 1, we assume that node N4 is the maliciousnode. Suppose node N1 wants to send data packets to© 2010 ACADEMY PUBLISHER

98 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Figure 2 : Data flow between Node 1 to Node 2 via Node 3 and Node 4Figure 3 : Link breakage and Data LossFigure 4 : Route Discovery ProcessTable 1: Simulation ParametersSimulatorns-2 (ver.2.31)Simulation Time500(s)Number of Mobile Nodes 20Number of Blackhole Nodes 1Topology750m x750mTransmission Range250mRouting ProtocolAODVTrafficConstant Bit Rate (CBR)Pause Time10(s)Maximum Connections 9Packet Size512 bytesData Rates10 KbitsWe have taken four scenarios of defined parametersfor our simulation with or without blackhole node. Wehave taken different positions and movements of nodesfor each scenario. Then we have varied the blackholenodes and simple nodes to evaluate the performance. Wehave also varied the mobility speed of mobile nodes.Themetrics are used to evaluate the performance are packetloss percentage, throughput and end-to-end delay. Wecalculate data loss percentage with blackhole and withoutblackhole node. Then we compare the results of thesetwo simulations to understand the network and nodebehaviors. The results of the simulation show that thepacket loss in the network with a blackhole increasesbeyond that dropped by the blackhole node. This is due toincreased congestion in the routes toward the blackholenode.Table 2 shows the packet loss percentage on AODVwith the presence and absence of blackhole nodes forfour scenarios.Table 2: Simulation Results with blackhole effect and withoutblackhole effectScenarios Total Loss Total Loss Increaseof AODV of BlackHoleAODVScenario1 3.37 90.54 87.17Scenario2 2.53 90.42 87.89Scenario3 2.35 98.54 96.19Scenario4 1.74 88.01 86.27Figure 5 : Node 1 found new route and Node 6(Blackhole Node)absorbs the DataOur simulation results show that AODV network hasnormally 2.50 % data loss and if a blackhole node isintroducing in this network data loss is increased to 89.38%. As 2.50 % data loss already exists in this data traffic,blackhole node increases this data loss by 86.88 %.We have also analyzed the throughput of receivedpackets with the presence and absence of blackhole nodewith respect to the simulation time of 450(s).Fig. 6 illustrates the graphic representation of packetloss percentage with and without blackhole node withrespect to simulation time (1=100seconds).© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 99Figure 7: Impact of Blackhole Node on Throughput of ReceivedPacketsFigure 6 : Shows the Packet Loss of AODV and blackholeAODVFig. 7 shows the effect of blackhole attack onthroughput of received packets of network. The resultshows both the cases with blackhole and withoutblackhole attack. With our simulation, we analysed thatthe throughput of received packets in AODV is very highthan the throughput of received packets inblackholeAODV. Because the packet loss inblackholeAODV is higher than the AODV protocol.We studied the performance with varying number ofBlackhole Nodes. Number of Blackhole Nodes variesfrom 1 to 4 with the increment of 1. Fig. 8 shows theimpact of number of Blackhole Nodes on throughput inthe network.. Simulation results show that the throughputdecreases with the increase of number of BlackholeNodes.We also studied the performance with varying thenumber of nodes. Fig. 9 shows the impact of number ofnodes on throughput without blackhole attack. Thenumber of nodes is varying from 10 to 50 with the step of10. Simulation results show that when the number ofnodes increases, the throughput increases for AODVprotocol.We have evaluated the End-to-End Delay with varyingthe mobility speed of nodes without blackhole node andwith blackhole node. The mobility speed varies from10m/s to 50m/s with the increase of 10.Fig. 10 illustrates the End-to-End Delay withblackhole attack and without blackhole attack. Weobserved that, there is increase in the average end-to-enddelay without the effect of blackhole attack. This is dueto the immediate reply from the malicious node becauseit doesn’t check its routing table.Figure 8: Impack of Number of Blackhole Nodes on ThroughputFigure 9: Impact of Number of Nodes on Throughput© 2010 ACADEMY PUBLISHER

100 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010REFERENCESFigure 10: Impact of blackhole attack on End-to-End DelayV. CONCLUSIONIn this paper, we studied AODV in detail and theblackhole attack in AODV. We evaluate the effects ofblackhole nodes on AODV in ad hoc networks. Wesimulate the blackhole behavior with the help of NetworkSimulator 2 and compared the performance ofblackholeAODV with the original AODV in terms ofpacket loss percentage. The simulation results show thatthe packet loss increases in the network with a blackholenode. Simulation results also show that the throughput ofthe network is decreased with blackhole attack ascompared to without blackhole attack.. When the numberof blackhole nodes increases the throughput decreases.We observed that the End-to-end Delay withoutblackhole attack is slightly increased as compared to theeffect of blackhole attack. The detection of blackhole inad hoc networks is still to be a challenging taskACKNOWLEDGEMENTI would like to take the opportunity to thank peoplewho guided and supported me during this process.Without their contributions, this project would not havebeen possible. I have a great pleasure in expressing mydeep sense of gratitude and indebtedness to Mr. JagpreetSingh, Lecturer, Teg Bahadur Khalsa Institute ofEngineering and Technology, Malout and Ms. Rajkumari,lecturer, University Institute of Engineering andTechnology, Panjab University, Chandigarh, mysupervisor for their continuous guidance and invaluablesuggestions at all the time during the research work. Myspecial thanks to all my friends for being supportive inmy hours of need. Finally, I would not forget to thank myparents for their love and blessings that support andencouraged me at every moment I need. They were thefirst ones that introduced the amazing world to me andencouraged me to explore the wonderful nature.[1] Latha Tamilselvan and Dr. V.Sankaranarayanan,“Solution to Prevent Rushing Attack in Wireless Mobile Adhoc Networks”.[2] Latha Tamilselvan, Dr.V Sankaranarayanan, “Prevention ofBlackhole Attack in MANET”. The 2 nd InternationalConference on Wireles Broadband and Ultra WidebandCommunications (AusWireless 2007) India, 2007 IEEE.[3] Mohammad O. Pervaiz, Mihaela Cardei, and Jie Wu,“Routing Security in Ad Hoc wireless Networks”, NetworkSecurity, 2005 Springer.[4] Elizabeth M. Royer, and Chai-Keong Toh, “A Review ofCurrent Routing Protocols for Ad Hoc Mobile WirelessNetworks,” IEEE Personal Communications, pp. 46-55,April 1999.[5] Satoshi Kurosawa, Hidehisa Nakayama, Nei Kato, AbbasJamalipour, and Yoshiaki Nemoto. “Detecting BlackholeAttack on AODV based Mobile Ad hoc networks byDynamic Learning Method”. International Journal ofNetwork Security, Vol.5, No.3, PP.338–346, Nov 2007.[6] ns-2 : http://www.isi.edu/nsnam/ns/[7] C. E. Perkins and E. M. Royer, “Ad Hoc On-DemandDistance Vector Routing,” Proc. 2nd IEEE Wksp. MobileComp. Sys. and Apps., New Orleans, LA, Feb. 1999, pp.90–100.Anu Bala from Bhogpur (Jalandhar, India)and her date of birth is 22 nd August 1985. She isa ME (IT) student in Information Technologyat the University Institute of Engineering andTechnology, Panjab University, Chandigarh,India. She received her B.Tech (CSE) degreefrom Panjab Technical University, jalandhar,India in 2006. Her research interests in the area of mobile adhoc networks, especially ad hoc network security. Her recentwork has focused on ad hoc routing protocol attacks.She has worked as Lecturer for one year at IITT college ofengg., Pojewal, Punjab, India.Raj Kumari from chandigarh (India) and his date of birth is6 th June 1981. she received her M.Tech (IT) degree fromGNDU, Amritsar, India in 2006 and her B.Tech from PanjabTechnical University, Jalandhar, India in 2003.She has works as Lecture for one year at college of Engg.Tangori, India. She has been working in UIET, Chandigarh,India since 2007Jagpreet Singh from Malout (India) and his date of birth is5 th March 1983. He received his MS (Software System) degreefrom BITS Pilani, India and his B.Tech from Panjab TechnicalUniversity, Jalandhar, India. He has been engaged in researchon mobile adhoc network and he has enhanced algorithm ofAODV protocol. He has been working in GTBIET, Malout,India since 2003. Two of his papers has been published inNational Conferences.1) Communication technology on UBI Quietestcomputing2) Successful implementation of recruisiteimplementation of e-governance© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 101Copyright Protection of Gray Scale Images byWatermarking Technique Using (N, N) SecretSharing SchemeSushma YalamanchiliProfessor & Head, Department of CSE,V.R.Siddhartha Engineering College, Vijayawada.Email: sushma_yalamanchili@yahoo.co.inM.Kameswara RaoLecturer, P.B.Siddhartha College, P.G.Centre, Vijayawada.Email: kamesh.manchiraju@gmail.comAbstract— Since the rise of the Internet one of the mostimportant factors of information technology andcommunication has been the security of information. Aspecial case of information hiding is digital watermarking.Digital watermarking is the process of embeddinginformation into digital multimedia content such that theinformation (the watermark) can later be extracted ordetected for a variety of purposes including copy preventionand control. Among numerous cryptographic solutionsproposed in the past few years, secret sharing schemes havebeen found sufficiently secure to facilitate distributed trustand shared control in various communication applications.In this paper, a new image watermarking algorithm isdeveloped using (n,n) secret sharing scheme for copyrightprotection. The proposed method embeds the copyrightimage into original image and is to be shared among nparticipants. Then the copyright image could be recoveredusing simple XOR operations without any loss.Experimental simulations are provided using MATLAB 7.1to demonstrate the efficient performance of the developedtechnique in terms of reliability of watermark embeddingand extraction. The scheme contains three phases: thecopyright image embedding phase, embedded image secretsharing phase, copyright image extraction phase. Theexperimental results show that the proposed scheme canresist several attacks such as JPEG compression, resize andnoise addition.Index Terms— copyright protection, secret sharing,watermarking, steganography.I. INTRODUCTIONOn the Internet today it is possible to duplicate digitalinformation a million-fold and distribute it over the entireworld in seconds. These issues worry creators ofintellectual property to the point that they do not evenconsider to publish on the Internet. To solve the problemof publishing digital images, researchers have come upwith digital image watermarking [1]. Digitalwatermarking is a method of embedding identifyinginformation in an image, in such a manner that it cannoteasily be removed .An application of watermarking iscopyright control, in which an image owner seeks toprevent illegal copying of the image. A digital watermarkis a code that is embedded inside an image. It acts as adigital signature, giving the image a sense of ownershipor authenticity. Digital watermarking of an image hasalso been proposed for the prevention of copying of animage by unauthorized persons.To a computer, an image is a collection of numbersthat constitute different light intensities in different areasof the image [4]. This numeric representation forms agrid and the individual points are referred to as pixels.Most images on the Internet consists of a rectangular mapof the image’s pixels (represented as bits) where eachpixel is located and its colour [5]. Digital colour imagesare typically stored in 24-bit files and use the RGB colourmodel, also known as true colour [6]. All colourvariations for the pixels of a 24-bit image are derivedfrom three primary colours: red, green and blue, and eachprimary colour is represented by 8 bits [4].Secret sharing [2] refers to method for distributing asecret amongst a group of participants, each of which isallocated a share of the secret. The secret can bereconstructed only when the shares are combinedtogether; individual shares are of no use on their own.Secret image sharing is a technique for protecting imagesthat involves the dispersion of the secret image into manyshadow images. This endows the method with a highertolerance against data corruption or loss than otherimage-protection mechanisms, such as encryption orsteganography.A secret kept in a single information-carrier could beeasily lost or damaged. Secret sharing (SS) schemes,called (n, n) schemes, have been proposed since late1970s. To encode the secret into n pieces (“shadows” or“shares”) that the pieces can be distributed to nparticipants at different locations. The secret can only bereconstructed from n pieces [2].II. PROPOSED METHODConsider an image A of NR ×NC. Each pixel of Acan take any one of c different colors or gray-levels.Image A is represented by an integer matrix A:A = [aij ]NR×NC , where i = 1, 2, . . . , NR,j = 1, 2, . . . , Nc, and aij {0, 1, . . . , c − 1}.We have c=2 for a binary image, and c =256 for a grayscale image© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.101-105

102 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010with one byte per pixel. In a color image with one byteper pixel, the pixel value can be an index to a color table,thus c = 256. In a color image using an RGB model, eachpixel has three integers: R (red), G (green) and B (blue).If each R, G or B takes value between 0 and 255, we havec = 256 [2]. The proposed method includes the followingstepsStep 1: Consider two images - Original image and thecopyright image - represented by integer matrices.Step 2: Decide a value called weight_value between 0and 1.For invisible watermark weight_value must be nearto 0.Watermark Embedding :Step 3: Embed the copyright image inside the originalimage as followsEmbedded image=original image + (resized copyrightimage * weight_value).(N,N) Secret Sharing Scheme For the embedded Image :The output of a (n,n) scheme for embedded images isa set of n distinct NR×NC matrices A1, . . . , An, calledshares or shadow images. Each share image has the samenumber of pixels as the original image A, but every pixelin a share may contain m sub pixels. A can bereconstructed from the set {Ai1, . . . , Aik } and evencomplete knowledge of k − 1 shares reveals noinformation about A. The first condition above is calledprecision, and the second condition is called security [2].Step 4: Encode the embedded image into n shadows orsecrets as follows andgenerate n − 1 random matrices B1, . . . , Bn−1,compute the shadow images as below:A1 = B1,A2 = B1 ⊕ B2,. . . . . .An−1 = Bn−2 ⊕ Bn−1,An = Bn−1 ⊕ A.In the generation of the shadow images and in thereconstruction of the secret, Boolean operation XOR(“ ⊕ ”) is used. For easy lookup, the truth-table of XORfor binary scalar inputs is given below.a=0 a=1b=0 0 1b=1 1 0a ⊕ bFor example, when a =125 and b =18, the XOR betweenthese two integers isa ⊕ b = (125)10 ⊕ (18)10 = (01111101)2 ⊕(00010010)2= (01101111)2 = (111)10.Step 5 : Reveal the embedded image from the n shadowimages as below:A’ = A1 ⊕ A2 ⊕ · · · ⊕ An.Because the “ ⊕ ” operation is associative and Bi ⊕ Biis a zero matrix for any i, we haveA = B1 ⊕ (B1 ⊕ B2) ⊕ · · · ⊕ (Bn−2 ⊕ Bn−1) ⊕(Bn−1 ⊕ A)= (B1 ⊕ B1) ⊕ · · · ⊕ (Bn−1 ⊕ Bn−1) ⊕ A = A.To demonstrate the computation steps in the revealingprocess, we give a trivial example for n = 3 and a singlepixelsecret image A.Given: A = (231)10 = (11100111)2Generate: B1 = (46)10 = (00101110)2 andB2 = (188)10 = (10111100)2Compute: A1 = B1 = (46)10 = (00101110)2,A2 = B1 ⊕ B2 = (10010010)2, andA3 = B2 ⊕ A = (01011011)2Reconstruct: A1 ⊕ A2 ⊕ A3 = (11100111)2 = (231)10.An example for the proposed (n, n) scheme using a 3× 3 secret image A is given below:A = 209 214 225233 227 228222 221 226B1 = 136 169 28254 128 24596 108 49B2 = 8 94 65218 137 22846 71 222,A1 = B1 =,A2 = B1 ⊕ B2 =,A3 = B2 ⊕ A =, andA1 ⊕ A2 ⊕ A3 =136 169 28254 128 24596 108 49128 247 9336 9 1778 43 239217 136 16051 106 0240 154 60209 214 225233 227 228222 221 226= AWatermark Extraction :Step 6 : Extract the copyright image inside the embeddedimage as followscopyright image =(Original image – embeddedimage)/weight_value.III. EXPERIMENTAL RESULTS ON GRAY SCALE IMAGESThe simulations were conducted on Intel machinewith 2.4 GHz processor and 512 MB of RAM. MATLAB7.1 was used for implementation of proposed scheme and© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 103image processing operations respectively. Applying theproposed method on tree path image as original imageand college logo image as copyright image under (n,n)scheme with n = 3.Figure 2(b)shadow image1Figure 1 (a) Original imageFigure 1 (b) Copyright ImageFigure 2( c ) Shadow image 2Figure 1(c) Embedded ImageFig. 1 shows the original image in part (a), copyrightimage in part (b) and the embedded image (Originalimage + Copyright Image) in part(c) .(n,n) Secret sharing schemeA (n,n) secret sharing scheme encodes a secret imageinto n share images, which demonstrate randomly noisypatterns and hide the information about the secret image,are distributed to n recipients. The secret image can easilybe decrypted by XOR operation of the n share images inan arbitrary order without complicated arithmetic.Figure 2(a)Embedded ImageFigure 2(d) Shadow image 3Figure 2 (e) Reconstructed Embedded image using 3shadowsFig. 2(a) shows the embedded image in part (a), shadowimages in part(b-d)and the reconstructed embeddedimage using the 3 shadow images in part(e).Strength of the (n,n) SchemeThe reconstruction phase of our algorithm computes nshadow images using XOR operation. the proposed (n, n)scheme reconstructs the secret image exactly, and itsatisfies the security condition. That is, when fewer thann shadows are used, the original secret image A will notbe revealed.© 2010 ACADEMY PUBLISHER

104 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Our experiments show that for group of attacks likeJPEG compression, resizing, adding noise the proposedmethod can be able to extract the copyright image withless loss in the quality.Figure 3(a)Reconstructed image using shadowimage 1 and shadow image 2Figure 5(a) Resized Embedded ImageFigure 3(b)Reconstructed image using shadowimage 1 and shadow image 3Figure 5(b)Extracted Copyright image when the embeddedimage is resizedFig 5 shows the Resized embedded image in part(a) andthe extracted copyright image when the embedded imageis resized in part(b)Figure 3(c)Reconstructed image using shadowimage 2 and shadow image 3Fig 3 shows the Reconstructed image using shadowimages 2(a) and 2(b) in part(a) Reconstructed imageusing shadow images 2(a) and 2(c) in part(b) andReconstructed image using shadow images 2(a) and 2(c)in part(c).Watermark extraction processFigure 6(a) Compressed Embedded ImageFigure 6(b)Extracted Copyright image when the embeddedimage is CompressedFig 6 shows the compressed embedded image in part (a)and the extracted copyright image when the embeddedimage is compressed in part (b)Figure 4(a) Embedded imageFigure 7(a) Embedded Image with added noiseFigure 4(b)Extracted copyright imageFig 4 shows the embedded image in part (a) and theextracted copyright image in part (b)Attacks on the Proposed MethodFigure 7(b)Extracted Copyright image when the embeddedimage is added with noise© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 105Fig 7 shows the embedded image with added noise inpart (a) and the extracted copyright image when theembedded image is added with noise in part (b)IV. PSNR MEASUREMENTOne commonly used measure to evaluate theimperceptibility of the watermarked image is the peaksignal to noise ratio (PSNR) which is given byPSNR = 10 * log10( ( 255 )^2 / mean_square_error )WhereOriginalImage_x_size = size( OriginalImage, 2);OriginalImage_y_size = size( OriginalImage, 1);CopyrightImage_x_size = size( CopyrightImage, 2);CopyrightImage_y_size = size( CopyrightImage, 1);mean_square_error = sum( sum( sum( ( CopyrightImage– Extracted CopyrightImage ).^2 ) ) ) / double(CopyrightImage_x_size * CopyrightImage_y_size * 3);[4]Johnson, N.F. & Jajodia, S., “Exploring Steganography:Seeing the Unseen”, Computer Journal, February 1998 .[5]“Reference guide: Graphics Technical Options andDecisions”, http://www.devx.com/projectcool/Article/1997[6] Owens, M., “A discussion of covert channels andsteganography”, SANS Institute, 2002.TABLE I.PSNR VALUESType of Operation on ImagePSNROriginal Embedded Image 32.6737After resizing embedded Image 29.0508After compressing the Embedded image to a .jpg 28.4083After adding random noise to the Embedded image 33.4338Table I illustrates PSNR values between the OriginalCopyright Image and the Extracted Copyright Imagetaken from various operationsV. CONCLUSIONSWe have demonstrated a new watermarkingtechnique that uses (n,n) secret sharing scheme to embeda copyright image into original image . This techniqueworks well with images of all sizes. This techniqueprovides two layers of security. In the first step, acopyright image is embedded into original image forcopyright protection. Also, the embedded image is sharedamong n participants where all the n shares must be usedto reconstruct the embedded image. This makes thesystem more secure. The method can with stand attackslike JPEG compression, resize and adding noise with lessloss in quality of the image. Further this work can beextended by calculating the hash value of the image andencrypt the hash value using either symmetric key orpublic key and embed the hash value inside the image sothat at the receivers end the authentication and integrityof the image can be verified by recalculating the hashand verifying it. Similarly digital signatures can begenerated for images and can be verified. Also (k,n)threshold secret sharing schemes can be implemented formuch security.VI. REFERENCES[1] Adrian Perrig Andrew Willmott, ”Digital ImageWatermarking in the Real World" Extended Abstract, March 9,1998.[2] DaoshunWang, Lei Zhang, Ning Ma, Xiaobo Li,“Two secretsharing schemes based on Boolean operations”, Science Direct–Pattern Recognition 2007.[3] A. Shamir, “How to share a secret”, Commun. Assoc.Comput. Mach. 22 (11) (1979) 612–613.© 2010 ACADEMY PUBLISHER

106 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Broadband Integrated Services over ProposedOpen CPE ArchitectureKushal RoyPrincipal Investigator to DST Government of India and faculty,Department of Electronics and Communication EngineeringHaldia Institute of Technology, ICARE Complex, PO H.I.T HaldiaPIN- 721657, West Bengal, IndiaEmail: profkushalroy@hotmail.comAbstract— Over the decades, due to technology limitationsand regulatory restrictions, information services have beendelivered to subscribers through multiple service providers.In recent years, however, the advancement of digitalcommunication technology, the passing of the 1996Telecommunications Act in the U.S., and the emergence ofthe Internet are driving network convergence. As a result,the industry is calling for the consolidation of an integratedCustomer Premises Equipment (CPE) in the customer’spremises to provide integrated services, such as multimedia,voice, and data services. However, the ever-changingnetwork standards and competing technologies not onlyconfuse carriers, but also prevent them from committing tomassive CPE deployment.In this paper, the problems carriers face in the deploymentof integrated services over broadband are first described. Inorder to remove these deployment obstacles, and also tocreate new value-added services, service models areinvestigated and an open CPE architecture is proposed thatwill support the wide range of broadband accesstechnologies and ever-changing network standards.Index Terms—Integrated Services, Broadband, VoIP, PSN,SoftSwitch Model.I. INTRODUCTIONThe convergence of voice and Internet data trafficalso calls for new Customer Premises Equipment (CPE)such as Digital Subscriber Lines (DSLs), modems, cablemodems, and wireless modems that provide users withbroadband access to the Internet. However, instead ofadding new CPE to the home, the industry is movingtowards consolidating CPE into an Integrated CPE (I-CPE) to help lower the cost, reduce network managementcomplexities, and improve the efficiency of networkresources. The I-CPE is also called an Integrated AccessDevice (IAD) and a Residential Gateway (RG) whenused in residential areas. In this paper, the term I-CPE isused, because I-CPE is intended to serve multiple marketsegments including residential, Small Office HomeOffice (SOHO), and small businesses to provideintegrated services such as voice, multi-media, andInternet access.While the upgrade of the network infrastructure iswell underway, the deployment of I-CPE remains a bigobstacle to the delivery of integrated services overbroadband due to the volume and time that are requiredfor the deployment. It was believed that the lack of autoconfigurationwas the main roadblock to massive I-CPEdeployment. However, the largest obstacle today is not somuch how to deploy I-CPE but rather, which type of I-CPE the carriers should be deploying, because thistelecommunication world is filled with too manystandards, e.g., Media Gateway Control Protocol(MGCP), H.248 [7], H.323, Session Initiation Protocol(SIP), Voice over Internet Protocol (VoIP), Voice overATM (VoATM), etc.; and competing technologies, e.g.,DSL, Cable, Wireless, Fiber, Ethernet, HomeRF ∗ , IEEE802.11, Bluetooth, etc. As a result, carriers are facinggreat difficulties in choosing an I-CPE for deploymentbecause they fear it being replaced in the near future.II. THE NEXT-GENERATION NETWORKINFRASTRUCTUREUnlike the dial-up modem, the introduction of I-CPEwill require major upgrades in the infrastructure of theexisting network. The new network is intended to supportthe convergence of the voice-centric Public SwitchedTelephone Network (PSTN) and the Internet, and it iscommonly referred as the Next-Generation Network(NGN). The NGN is based on the distributed Softswitcharchitecture in which the call control is separated fromthe media transport. Figure 1 shows the role of I-CPE inthe NGN infrastructure. It indicates that an I-CPE is theportal to a customer’s premises and it acts as the gatewayto interconnect the Local Area Network (LAN) with theWide Area Network (WAN), which consist of the AccessNetwork and the Core Network [6]. The Access Network,consisting of Access Nodes, Regional BroadbandNetworks, Transit Gateways, and Because of the passingof the Telecommunications Act in 1996 and the FCC’songoing adventures in local loop deregulation,companies, including Inter-eXchange Carriers (IXC),Incumbent Local Exchange Carriers (ILEC), CompetitiveLocal Exchange Carriers (CLEC), CATV serviceproviders, and many other emerging service providers,are all eager to enter this lucrative broadband accessbusiness. Therefore, it is certain that I-CPE will supportvarious broadband interfaces, such as wireless, cable,xDSL, and fiber optics, to target different market sectorsdepending on price, performance, environment, or thetype of users. Then, the recent advancement of IEEE© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.106-109

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 10780.3ae 10 GigE standard [13] enables the Ethernet towork beyond the LAN. Thus, an I-CPE may also supportEthernet broadband interface to the WAN. Access Nodesterminate the broadband interfaces from I-CPE andaggregate multiple CPE traffic to the RegionalBroadband Networks. Regional Broadband Networksprovide transport and switching functions among AccessNodes and Transit Gateways. Transit Gateways mayconsist of Trunking Gateways and Signaling Gatewaysthat perform signaling and bearer interworking functionsbetween Regional Broadband Networks and PSN/CSN.An MGC is also referred to as the Call Manager orSoftswitch that typically runs call control software in aserver to provide call-processing functions. In thedistributed Softswitch architecture, the value-addedservices or applications can reside in an ApplicationServer. A Media Server is controlled by the Softswitch toplay tonal announcements or perform media streamingfunctions, such as Interactive Voice Response (IVR) andConference Bridge. The media streaming is carried inbearer connections between the Access Node and theTransit Gateway for on-net calls, or between AccessNodes for off-net calls, while call control functions areprovided by an MGC. GigaPoP (i.e., Point of Presence)aggregates the traffic from Ethernet CPE to PSN [14]. Itshould be noted that the NGN architecture shown aboveis meant to be logical, so the Access Nodes, TransitGateways, Media Server, Application Server, and MGCare logical functional blocks which may be implementedas stand-alone devices or embedded in other networkelements.The Media Gateway Controller (MGC), provides theramp to Packet-Switched Networks (PSN) and Circuit-Switched Networks (CSN).Figure 1: Next-Generation Network infrastructureIII OPEN CPE ARCHITECTUREBroadband access is, for many, the solution to the“World Wide Wait” problem (slow Web performance),and it would also enable the creation of many newservices that promise enormous benefits to subscribersand service providers. However, service providers arefacing many challenges during the deployment ofbroadband access networks, not least of which is how todeploy the greatest number of Customer PremisesEquipment (CPE) in the timeliest manner..A. I-CPE CharacteristicsThe following characteristics of the IntegratedCustomer Premises Equipment (I-CPE) architecture areimportant to the resolution of these deployment problems.• Open Architecture–In the past several decades,open architecture has become the trend in theindustry. It started with the computer industry inthe early 80s, and it gradually spread to thetelecommunication industry, where the monolithiccentral office switch is under great pressure fromthe Internet telephony to open up. The broadbandCPE industry will not buck this trend sinceopenness will only foster new services and productcompetition, which, in turn, will lower CPE cost,reduce time-to-market, and inspire innovation [3].• Flexibility–The ever-changing network standardsand infrastructures, the wide range of accessnetwork technologies, and the uncertainty ofwhich value-added service will emerge in thefuture make it very difficult to have a fit-it-all CPEarchitecture. Thus, the I-CPE should be flexible tocontain the must-have core features needed tosupport today’s services, yet it should still be ableto accommodate future upgrades. Flexiblearchitecture will lower CPE costs, something thatis crucial to I-CPE being accepted in the costsensitiveconsumer market. For example, an I-CPEmight be designed to provide home automationand security services, but the interface standardsas well as the functionalities to control homeappliances have yet to be defined. The flexiblearchitecture needs to be able to accommodate aBluetooth wireless modem plug-in and thenecessary software that will need to be added tothe CPE to support this feature when the standardand technology become available.• Value-Added Services - Each time a new serviceor technology is introduced, it needs to beperceived by the public as adding value in orderthat it will be bought and used. Due to thecompetition from Internet telephony, the cost oftoll telephony services is continually decreasing.Soon, it is believed, you will be charged a flatmonthly rate for telephone services instead ofbeing billed on a per-minute basis. Therefore, thedriver for the Next-Generation Network (NGN) isnot about inventing a new way to provide existingservices, but focusing on value-added servicessuch as Presence, Voice Web, Voice Portal,Unified Messaging, and Voice Virtual PrivateNetwork (VPN), which hold great potential forgenerating additional revenue for serviceproviders.© 2010 ACADEMY PUBLISHER

108 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010B. Integrated CPE Functional DecompositionFigure 2 shows the Open CPE architecture that isintended to fulfill the requirements as listed above and tosupport multiple broadband access technologies andtelephony protocol standards. It is composed of a CPEController Module (CCM), LAN Interface Modules(LIM), and WAN Interface Modules (WIM). The CCMconsists of a CPE Controller, Flash, ROM, and RAM thatcontain a Real-Time Operating System (RTOS) andsoftware to implement telephony call control, networkmanagement, and service logic functions, as well as thirdpartyapplications. The CCM may also include a DigitalSignaling Processor (DSP) to implement media streamingprocessing functions, such as vocoder, echo cancellation,tone generation and detection. The CCM can be reconfiguredto support different protocols or telephonystandards through software downloading.Figure 2: Open CPE architectureWIM and LIM contain plug-ins to support variousWAN and LAN interfaces. WIM may consist of aWireless Modem , a Cable Modem, a DSL Modem, aFiber Modem, and a 10 Gigabit Ethernet module thatprovide interfaces to the broadband access network. LIMmay include interfaces to analog phones. There are twobuses responsible for the distribution of user data, realtimevoice streaming, control signaling, and themanagement data between the CCM Controller and theWIM/LIM. The interface specification should meet therequirements of transporting real-time and non-real-timedata. The open CPE architecture as described above isderived from the abstraction of multiple CPE platformsthat may use various broadband access technologies andstandards. The goal of the abstraction is to find out whatfunctions are common to all CPEs and therefore shouldbe implemented in CCM, and what functions are interfacedependent. and therefore are more appropriatelyimplemented in plug-in modules. Figure 3 gives anexample of the CPE functional decomposition. It depictsthe protocol stacks of I-CPE supporting cable, xDSL, and10 Gigabit Ethernet broadband interfaces to provide voiceand data convergence services.Figure 3: I-CPE protocol stackThe Cable CPE is intended to provide VoIP services.It uses Network-based Call Signaling (NCS) [4] as thesignaling protocol to perform telephony call controlfunctions, while encapsulating the voice streaming data inthe Real-Time Protocol (RTP) packets to be sent to theCable Modem Termination System (CMTS). xDSL CPEprovides VoIP and Voice over AAL2 (ATM AdaptationLayer type 2) Loop Emulation Services (LES) [5]. It usesH.248 [7] and Channel Associated Signaling (CAS)protocols to implement telephony call control functionsfor VoIP and VoAAL2, respectively. The voice streamingdata are encapsulated in either RTP or AAL2 packets tobe sent to the DSL Access Multiplex (DSLAM). In thisexample, the Ethernet CPE acts as a SIP client to provideVoIP services. SIP is used to establish, modify, andterminate multi-media sessions and calls [10]. The voicestreaming is encapsulated in the RTP packets transmittingto the terminating party via the User Datagram Protocol(UDP) connection.In the Time Division Multiplex (TDM) networks,only call control and management functions areimplemented by software in the processor, leaving thevoice processing to be handled by the hardware, becauseof the latency concern.However, the trend toward packet telephony, alongwith the advancement and increasing performance ofprocessors in recent years, has made it possible for, andeven demanded that, voice streaming be processed bysoftware. Thus, as Figure 3 indicates, it makes good senseto locate the CCM-WIM interface between Layer 2 andLayer 3 of the protocol stack. The CCM containssoftware to implement voice processing, the signalingprotocol, and management functions. WIM shouldimplement Physical layer (PHY), Medium AccessControl (MAC), Logical Link Control (LLC),Asynchronous Transfer Mode (ATM), and ATM.Adaptation Layer (AAL) functions that are closelycoupled with each broadband access interface.IV. CONCLUSIONThe explosion of the Internet along with regulationand technology changes are reshaping telecommunicationnetworks in many ways. The industry is moving towardthe convergence of PSTN and the Internet. The© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 109converged network, the NGN, will operate in a verysimilar way to the Internet topology in which centraloffice switches are decomposed into distributed systemsthat are based on the Softswitch model. As a result, theNGN will no longer provide transport services betweentelephone equipment (as PSTN previously did), but willprovide personalized multi-media services. Thisconvergence of voice and Internet data traffic also callsfor the deployment of I-CPE to provide integratedservices. However, the ever-changing network standardsand competing technologies not only confuse carriers, butalso prevent them from committing to massive CPEdeployment, because they are afraid that the CPE justdeployed will have to be replaced in a few years.Hardware replacement is very common in the PC andcellular phone business when new standards ortechnologies are introduced, but it presents a big threat tothe wireline broadband business because CPEdeployment is such a daunting task that it cannot becompleted easily. In this paper, I proposed an open CPEarchitecture to solve the CPE deployment dilemma. Thearchitecture is very flexible and can support multipleWAN/LAN technologies and IP Telephony standards. Italso includes a common API to allow users to customizeor even create new services and third-party developers tocreate new applications and services. The open CPEarchitecture will allow CPE vendors to lower costs andreduce time-to-market, and most importantly enableservice providers to provide many value-added servicesthat promise great potential for generating additionalrevenueREFERENCES[1] C. A. Eldering, “Customer Premises Equipment forResidential Broadband Gateway,” IEEE CommunicationMagazine, , pp. 114-121, June 1997.[2] S. Moyer and A. Umar, “The Impact of NetworkConvergence on Telecommunications Software,” IEEECommunication Magazine, pp. 78-84, January 2001[3] C. Low, “Integrating Communication Services,” IEEECommunication Magazine, pp. 164-169, June 1997.[4] “PacketCable Network-based Call Signaling ProtocolSpecification,” Pkt-SP-EC-MGCP-101-990312, Cable Lab,3/12/1999.[5] “Voice and Multimedia over ATM–Loop Emulation ServiceUsing AAL2,” af-vmoa-145.000, ATM Forum, July 2000.[6] J. Chou, “The migration of LES to the Next GenerationNetwork based on H.248,” ATM Forum, atm00-164, SanFrancisco, May 2000[7] F. Cuervo, N. Greene, A. Rayhan, C. Huitema, B. Rosen,and J. Segers. “Megaco Protocol Version 1.0,” IETFRFC3015, November 2000.[8] J. Lennox, H. Schulzrinne, “Call Processing LanguageFramework and Requirements,” IETF RFC2824, May 2000.[9] J. Rosenberg, J. Lennox, H. Schulzrinne, “ProgrammingInternet Telephony Services,” IEEE Internet Computing,May-June 1999, pp. 63-72.[10] M. Handley, H. Schulzrinne, E. Schooler, and J.Rosenberg, “SIP: Session Initiation Protocol,” IETF,RFC2543, March 1999.[11] “OpenCable,” http://www.opencable.com, Cable Labs.[12] J. Lennox, H. Schulzrinne, J. Rosenberg, “CommonGateway Interface for SIP,” IETF RFC3050, January 2001.[13] IEEE P802.3ae 10Gb/s Ethernet Task Force,http://grouper.ieee.org/groups/802/3/ae/index.html.[14] “Lighting Internet in the WAN,” TelecommunicationMagazine, September, 2000, http://telecomsmag.com/issues/200009/tcs/lighting_internet.html.Kushal Roy, born in M.P Republic of India on December13 th 1979, received his Bachelors Degree with Honors inElectronics and Communication Engineering form GovernmentEngineering College Ujjain, M.P India, in the year 2002,achieved his Master’s Degree M,Phil in Instrumentation fromIndian Institute of Technology Roorkee, U.A Republic of Indiain the year 2004 with nano scale instrumentation and optoinstrumentation as major fields of study.He is involved extensively involved in Research educationand development since 2004 with more than 5 years ofexperience in teaching undergraduate as well as post graduatestudents in various renowned Institutes across the country.Presently he is serving as faculty in the Department ofElectronics and communication Engineering, Haldia Institute ofTechnology ICARE (Indian Center for Advancement inResearch and Education) Complex, Haldia, West BengalRepublic of India.Mr. Roy has contributed to more than 6 internationalconferences papers across the globe in Italy, Singapore,Morocco, USA etc. He is also working as Principal Investigatorto a sponsored project on “Study and Development of nanoscale piezo electronic material Lead Zirconium Titanate,(popularly known as PZT) by sol-gel process.” (conferred videHonorable President of India Direct sanction orderSR/FTP/ETA-49/07) worth 0.522 million INR, Science andEngineering Research Council, Department of Science andTechnology, Ministry of Science and Technology Governmentof India.© 2010 ACADEMY PUBLISHER

110 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Recovery Based Architecture To Protect HidsLog Files Using Time StampsSurinder S. KhuranaPunjab Engg. College, Chandigarh, Indiasurindersingh.cs07@pec.edu.inDivya Bansal , Prof. Sanjeev SofatPunjab Engg. College, Chandigarh, IndiaAbstract – After the great revolution in the field ofInformation Technology, many applications made necessityto run computer systems (either servers or client machines)all the time. Along with improvements and new inventionsin technology, the threat of attacks through computernetworks becomes a large issue. Host Based IntrusionDetection is a part of security system that protects hostsfrom various kinds of attacks. It also provides a greatdegree of visibility (of system activities). It is quite widestthat HIDS are vulnerable to attacks. An adversary, ifsuccessfully enters in a system can disable HIDS or modifyHIDS rules to hide its existence. One can easily evade HIDS.In [7] we propose a new architecture that protects HIDSfrom such attacks. In this paper, we have proposed a newmechanism to check integrity of log files. We have discussedits affects on performance of system.I. INTRODUCTIONIntrusion Detection System [5] is an imperativeingredient in network computer security which plays avital role in detecting the intrusive activities before theyoccur. Intrusion Detection is usually done throughscanning network traffic and/or hosts data and activities.These intrusive activities can be defined as - activitiesperformed by some adversary for gaining unlawfulbenefits. The adverse affects of such intrusion activitiesare in terms of loss of confidentiality, integrity andavailability of resources or services.IDSs can be classified under various categories.Figure-1 illustrates the various classifications of IntrusionDetection Systems.Response BasedDomain OfDetection BasedUnderlyingDetectionTechniqueBased• Active IDS• Passive IDS• HIDS• NIDS• Anomaly based IDS• Signature based IDSFigure-1: IDS ClassificationAn IDS can be active or passive. Passive IDS detectthe attacks and logs information or raise alarms. ActiveIDS takes action in response to an already detectedattack. Active IDSs are also known as IntrusionPrevention System.Section 2 discusses Detection and Recovery basedArchitecture to protect HIDS. Section 3 describes timestamping based protocol to check integrity of log files. Insection 4 and 5, we discuss implementation details.Affects of our architecture on system performance hasdescribed in section 6. In section 7, we discuss somerelated works. We present directions for future work in 6and our conclusion in section 8.II. OVERVIEW OF DETECTION AND RECOVERYBASED ARCHITECTUREIn [1], architecture has proposed to detect attacks onHIDS and recover HIDS to its previous healthy state. Thearchitecture protects HIDS from two type of attacks: firstis that it does not allow adversary to kill the HIDSprocess. And the other is it does not allow unauthorizedmodification of rule or signature database. The basic ideabehind the architecture is to allow the adversary toperform attack on HIDS and then recover the HIDS to itsprevious healthy state. A new process called MonitorIDShas been introduced in the proposed architecture.MonitorIDS is a lightweight system process whichdetects the attacks that affects HIDS and takes requiredactions to recover HIDS from affects of that attack.MonitorIDS takes care of both HIDS process andintegrity of HIDS related information. However, thisarchitecture does not consider underlying technique usedby HIDS to detect intrusions. It can be used with eithersignature based or anomaly based IDS.Figure 2 depicts working of proposed architecture. Asshown in figure 2, a backup of HIDS related files hasbeen created immediately after the installation.MonitorIDS process is embedded with HIDS. Thisprocess regularly monitors the HIDS process and filesafter a small fraction of time gap. If MonitorIDS foundHIDS process dead (detect unauthorized kill of HIDS) itimmediately restarts the HIDS. It also monitors integrityof files related to HIDS. These files may include rule orsignature database. If it found any unauthorizedmodification of HIDS files it replace the modified files© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.110-114

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 111Create Backpup Directory During InstallationRun HIDS and Monitor IDS processMonitor for Attack on HIDSMonitor for Attack on MonitorIDSAttack performedAttack performedHIDS ProcessTerminatedHIDS FilesChangedMonitorIDS ProcessTerminatedMonitorIDS relatedFiles ChangedRestart The HIDSRestore Files FromBackup DirectoryRestart MonitorIDSProcessRestore Files FromBackup DirectoryCreate Log For Attack and inform the Administartor.Figure 2 : Block Diagram of proposed architecturewith files from backup directory. It also logs the attackinformation and informs the administrator.The architecture also takes care of backup directoryand MonitorIDS process related information, so that thesecannot be modified by some adversary. One importantpoint to note about MonitorIDS is that it is a Systemprocess. To run this process as system process an entrylabeled with respawn has made in /etc/inittab file. Thatcannot be stopped. If some adversary makes attempt tokill this process, operating system kernel automaticallyrestarts it. MonitorIDS process also takes care of/etc/inittab file so that entry related to it cannot beremoved.III. TIME STAMPING BASED MECHANISM TOCHECK INTEGRITY OF LOG FILESCompares Last ModificationTime of Log FileWith stored valueWrite log entry in log fileandin backup copy of log file.Encrypt Last ModificationTime of log fileandIts backup copyReplace Log File WithBackup coopyThe main goal is to detect and recover unauthorizedmodification of log files by any process that is not relatedto HIDS. Such detection is made based on the lastmodification time (a file attribute, changed when the filewas modified) of the log file.As shown in Figure-3, when HIDS is running underthis mechanism it should follow the below givensequence of steps to write an entry in a file.1. Write log entry in log file and in backup copy oflog file.2. Encrypt Last Modification Time of log file and itsbackup copy.3. Write encrypted version of last Modification timeof log file and its backup copy into a file.Save Encrypt Last ModificationTimes in a FileFigure - 3: Sequence of Steps to Write Entry in Log FileBefore writing any entry IDS process compares thecurrent last modification time of log file and lastmodification time that was stored in a file as specified instep 3 given above. A mismatch between these twovalues represents that any other process changed the logfile means unauthorized modification of log file. In sucha case log file is replaced with its backup copy or ifunauthorized access to the backup file is detected, it willbe replaced with log file.© 2010 ACADEMY PUBLISHER

112 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010IV. IMPLEMENTATIONThe proposed architecture has been implemented inRed Hat Linux environment. Our architecture does notemphasis on underlying technology used by host basedintrusion detection system to detect intrusion. Rather thanconstruct a HIDS from scratch, we decided to leverage anHIDS within current implementation of our architecture.The two hypotheses that underlie this dissertation arepractical in nature. First, they intend to show that it isfeasible to protect HIDS using proposed architecture.Second, proposed architecture does not affect systemperformance adversely. Therefore, an implementationwas a center point for the development of this dissertationand was used both for practical verification of theintended features of the architecture and for aiding inreasoning about and experimenting with itscharacteristics.V. NEW PROCESSES PROPOSED INARCHITECTUREPractically MonitorIDS process (introduced in ourproposed architecture) has implemented with two subprocesses :1. MonitorProcesses2. MonitorFilesTo ensure that MoitorIDS processes live forever werun these processes as system processes. When theadversary tries to kill this process kernel automaticallyrestarts it. A respawn labeled entry related to theseprocesses has been written in system file ‘\etc\inittab’ torun these processes as system processes.MonitorProcesses process ensures that all processesrelated to OSSEC are always running. If it found anyprocess dead it restarts the process corresponding HIDSprocess. MonitorProcesses also prevents modificationrelated to these entries in ‘\etc\inittab’ file. In case ofdeletion or modification of these entries from‘\etc\inittab’ file, this process writes new entries. Thepurpose of MonitorFiles sub-process is to protect HIDSfiles from unauthorized modification. If it detects anymodification or deletion, it restores the victim file withgenuine file from backup directory.VI. AFFECTS ON PERFORMANCE OF THESYSTEMTo evaluate the affects on the performance on systemfollowing steps are carried out :1. To evaluate the affects on execution time of basicLinux commands, some time measurementswere carried out.2. System Monitoring Utility was used to check thechanges in utilization of processor due toprocesses related to our proposed architecture.A. Affects on execution of Linux commandsFor this purpose, most of the time measurements werecarried out regarding basic Linux commands (ps, find,who). Each Command was executed 100 times and allcorresponding 100 execution times were recorded. Theaverage of these timings has been considered as theAverage Execution Time(ATE). Average Execution Time(AET) was calculated twice. In first run, we calculateexecution time (ATE1) without running processes (suchas MonitorIDS process) related to our architecture. Andin second run we calculate execution time (ATE2) whileprocesses related to our architecture were running.Table-1 Average Execution Time in millisecondsCommand ps –ef find / >/dev/nullNumber of systemcalls 536 10055(a) Average Execution Time(AET 1) 25.9 63.1(time required to executeon machine while not runningprocesses related to proposedarchitecture )(b)Average Execution Time(AET 2) 30.1 65.5(time required to executeon machine while processesrelated to proposed architectureare running )Overhead relative to (a) 0.16% 0.03%Table 1 represents average execution time (inmilliseconds) and corresponding overhead. Very smallfraction of time was observed as overhead due to theMonitorIDS process.B. Changes in processor utilization due MonitorIDSProcessMonitorIDS process is a very lightweight processthat works in iterative manner. In each iteration checksthe state (process terminated or running) of HIDSprocesses and integrity of HIDS related files. Oneiteration requires approximately .001 Second time forexecution.Figure 3 represents the CPU utilization graph whenMonitorIDS process was not running. Because theprocessor in use is Dual Core two lines representsutilizations of processor. As shown in the graph CPUutilization is between 0% to 20%. Most of the time theCPU utilization is below 10%.Figure 3: Graph showing CPU utilization while MonitorIDS not running© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 113But as shown in graph given in figure 4, whileMonitorIDS process has been running the CPU use isincreased to some extent. Most of the time one of theCPU utilization lies between 15 to 20 % and other CPUutilization lies between 5 to 15%.even more time for execution as in comparison to time itrequires to execute on a normal machine.Table-2 Virtual Machine Overhead on execution timeCommand ps –ef find / >/dev/nullNumber of system calls 536 10055(a) Host Time 25 125(time required to executeOn real machine )(b) Guest Time 68 484(time required to executeOn real machine )Overhead relative to (a) 172% 287%Figure 4 Graph showing CPU utilization while MonitorIDS running.As in comparison to graph in figure3 the CPU use isincreased about approximately 8 to 10% of total CPUutilization while MonitorIDS process has been running.VII. RELATED WORKTo protect the HIDS many researchers have given theproposals. In [2], Laureano and Jamhour define the use ofvirtual machine to protect the HIDS. Virtual machine isused because of its inherent properties like separation ofexecution space. Other benefits [3] of virtual machine,that are useful in system security are isolation (a processrunning outside the VM cannot be accessed internalprocess of VM), inspection (VM state can be accessed byVM Monitor), and interposition (any operation issued byVM can be modified by VM Monitor). The idea behindthe architecture is to encapsulate the system activities(both user and guest activities) in virtual machine andplace the Intrusion Detection System outside the scope ofvirtual machine. Now any process has not been able toaccess the IDS. IDS monitors the activities performed inthe virtual machine and identify the intrusive activities.Their approach use type II virtual machine. Thearchitecture protects the IDS from attacks by placing itout from the scope of other processes. However, thisapproach has one basic problem regarding withtheperformance issue. The overhead due to virtualmachine degrades the system performance to very worsestatus. As results given by them if we compare theexecution timings of basic routines under an actualmachine and virtual environment, there is a largevariation. The time taken by routines under virtualmachine is much more than time taken on an actualmachine.Table 2 represents a subset of results given byLaureano and Jamhour in [2]. The virtual machineoverhead is so high that each routine requires double orThe architecture also affects the networkperformance. The performance of applications such asFTP and HTML etc. is also turned down.The work in [4] represents the use of virtual machinetype-I. The basic idea is same as to run the HIDS inexecution space that cannot be accessible by otherprocesses. However, virtual machine overheads alsoaffect this architecture.The use of virtual machines for the security ofsystems has defined by G. Dunlap & et. al.[6]. Theproposal defines an intermediate layer between themonitor and the host system, called Revirt. This layercaptures the data sent through the syslog process (thestandard UNIX logging daemon) of the virtual machineand sends it to the host system for storing and lateranalysis. However, if the virtual system is compromised,the guest syslog process can be terminated and/or the logmessages can be manipulated. by the intruder, andconsequently they are no longer reliable.VIII. FUTURE WORKStill there are many issues to be addressed about howthe proposed architecture can be best implemented andused. MonitorIDS process checks HIDS continuouslyin iterative way. To check HIDS in an iterationMonitorIDS process requires .001 seconds.On an average, the adversary has .0005 (.0001/2)seconds to disable our architecture and HIDS byexecuting following steps :1. Terminate the MonitorIDS process2. Remove the entry related to MonitorIDS processfrom /etc/inittab file to stop the Operating systemto restart MonitorIDS process.3. Terminate HIDS processes or change files relatedto HIDS.A DFA (Deterministic Finite Automata) shown inFigure 4, represent such sequences of steps. State q1 isthe initial healthy state and q4 is the final state. Afterreaching at q4 state attacker can tamper the HIDS. Toavoid this case, there should be some mechanism that© 2010 ACADEMY PUBLISHER

114 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010prevents transitions of system states from q1 to q4. Asdiscussed above in average case transition from q1 to q4is only possible if made in .0005 seconds. Suchmechanism can be based on detecting system callsequence required for this transition. If any suchsequence is detected it should delay (approximately .0005execution of these system calls so that before the systemstate will change from q2 or q3 to state q4, transitionfrom q1 or q2 to state q1(initial healthy state) took place.This mechanism can be implemented by changingoperating system kernel.IX. CONCLUSIONIn this paper, we propose a new time stamping basedapproach to check integrity of log files of HIDS. Thisapproach can be combined with Detection and AutoRecovery based architecture to Protect Host Based IDS.As discussed in section 4, unlike other approaches ourapproach does not use virtual machine and hence doesnot affect system performance adversely.Our mechanism ensures that HIDS process alwayslive (cannot be killed by adversary) and the informationrelated to HIDS can never be updated by any adversary.This architecture also ensures the integrity of frequentlyupdated log files of HIDS.REFERENCESFigure-4: DFA representing sequence of steps to disable ourarchitecture.Issue of authorization and integrity of genuineupdates of HIDS information is another important issueto be addressed. In this approach, we did not consider anytechnique to authorize of update. Because IDS’srules/signature database requires frequent updates,inclusion of a strong authorization mechanism would berequired.[1] A. Abraham, C. Grosan and C.M. Vide. EvolutionaryDesign of Intrusion Detection Programs. InternationalJournal of Network Security, Vol. 4, No. 3, 2007.[2] M Laureano, C Maziero, E Jamhour, Protecting hostbasedintrusion detectors through virtual machines-Computer Networks- Elsevier, 2007.[3] P. Chen, B. Noble, When Virtual Is Better Than Real,Workshop on Hot Topics in Operating Systems, 2001.[4] T. Garfinkel, M. Rosenblum, A virtual machineintrospection based architecture for intrusion detection,ISOC Network and Distributed System SecuritySymposium (2003).[5] S. Axelsson. Research in intrusion detection systems: Asurvey. Technical report, Chalmers University ofTechnology, 1999.[6] G. Dunlap, S. King, S. Cinar, M. Basrai, P. Chen, ReVirt:Enabling Intrusion Analysis through Virtual-MachineLogging and Replay, USENIX Symposium on OperatingSystems Design and Implementation, 2002.[7] Surinder Singh khurana, Ms. Divya Bansal, Prof. SanjeevSofat “Detection and Auto Recovery Approach to ProtectHost Based IDS” 2009 IEEE International AdvanceComputing Conference (IACC 2009)© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 115Software RadioVarun SharmaInfosys Technologies Limited, Chandigarh, IndiaEmail: Varun_sharma15@infosys.comYadvinder Singh MannInfosys Technologies Limited, Chandigarh, IndiaEmail: YadvinderS_Mann@infosys.comAbstract— This paper aims to provide an overview onrapidly growing technology in the radio domain whichovercomes the drawbacks suffered by the conventionalanalog radio. This is the age of Software radio – thetechnology which tries to transform the hardware radiotransceivers into smart programmable devices which can fitinto various devices available in today’s rapidly evolvingwireless communication industry. This new technology hassome or the entire physical layer functions software defined.All of the waveform processing, including the physical layer,of a wireless device moves into the software. An idealSoftware Radio provides improved device flexibility,software portability, and reduced development costs. Thispaper tries to get into the details of all this. It takes onethrough a brief history of conventional radios, analyzes thedrawbacks and then focuses on the Software radio inovercoming these short comings.I. HISTORYRadio, as anyone would perceive, is a device that canwirelessly transmit or receive signals in the radiofrequency (RF) part of the electromagnetic spectrum. J.CMaxwell postulated the theory of electromagnetic wavepropagation which was confirmed by H Hertz. Theseelectromagnetic waves travel through space eitherdirectly, or have their path altered by reflection,refraction or diffraction. When EM Waves come incontact with a conductor; they get converted to anelectrical energy, radio or micro wave depending on itswavelength. Radio waves can carry information byvarying a combination of the amplitude, frequency andphase of the wave within a frequency band. Nikola Teslaand Guglielmo Marconi invented devices that used radiowaves for communication.Earlier, radio systems relied entirely on the energycollected by an antenna to produce signals for theoperator. Radios became more useful after the inventionof electronic devices such as the vacuum tube and laterthe transistor, which made it possible to amplify weaksignals.Today most of our gadgets contain a radio system in it.From a cell phone to television , a walkie-talkie in a toy,space vehicles, cell phones or Satellite phones,Audio/Video broadcasting, Navigation Systems, RADARand many other applications. Traditional Radios werebuilt only for a particular frequency range, modulationtype, and output power. “Figure 1. shows a typicaltraditional radio receiver.The RF signal is converted to Intermediate Frequency(IF) using a programmable local oscillator. The IF signalis a fixed frequency and the IF signal is then amplifiedand filtered before feeding it to demodulator. Each kindof radio will have its own type of demodulator to extractthe audio data for playing it on speaker. These radioswould require hardware changes to modify thesefundamental characteristics like frequency range,modulation type etc. Moreover, with the advancement oftechnology, new communication standards keep comingand are used to varied degree in different countries e.g.CDMA, GSM EDGE, 3G, etc. The Conventional radioscannot cope with these advances, due to compatibilityissues. Hence, there was a need for dynamicallyconfigurable radios which could be used for variousapplications by simply re-configuring the softwarerunning in them. They can also be made to comply withthe various communication standards, just by changingthe implemented protocol. This is the world of SoftwareRadio.Figure 1. Block Diagram of a Traditional Radio Receiver© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.115-121

116 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010II. WHAT IS A SOFTWARE RADIO?Joe Mitola, who came up with the term Software Radiodefines it as "A radio whose channel modulationwaveforms are defined in software. That is, waveformsare generated as sampled digital signals, converted fromdigital to analog via a wideband DAC and then possiblyup converted from IF to RF. The receiver, similarly,employs a wideband Analog to Digital Converter (ADC)that captures all of the channels of the software radionode. The receiver then extracts, down converts anddemodulates the channel waveform using software on ageneral purpose processor." [2]The software radio, by his definition, should have aslittle hardware as possible.A Software Radio consists of a receiver and atransmitter. For the reception case as shown in “Figure 2.” [1], when the Antenna receives the signal, it passes thesignal to the receiver section for filtering and separatingthe signal from noise and interference. The signal is thenconverted to a desired frequency and amplitude which iscompatible with the analog to digital converter (ADC),and finally to digital data. The digitized data is thenprocessed using digital signal processing techniques forfurther use.For the transmission process as shown in "Figure 3.[1], the software programmable digital signal processingtechniques are used for generating the modulatedsignal(s) in the digital domain. The digital samples arethen converted to analog signal using Digital to AnalogConverter (DAC) and then amplified to the appropriatevoltage, current or power before transmission through theantenna.III. COMPONENTS OF A SOFTWARE RADIOSoftware Radio provides a flexible radio architecturethat makes it re-programmable for different usages withno or minimal change in the hardware. This is supportedby features like interference rejection techniques,encryption, voice recognition and compression, softwareenabledpower minimization and control, differentaddressing protocols and advanced error recoveryschemes.The technology provides the flexibility to combinedifferent grades of hardware and software to strike theright balance between cost and network resilience. Thereconfigurable blocks in Software Radio allows easychanges to the radio’s fundamental characteristics such asmodulation types, operating frequencies, bandwidth,multiple access schemes, source and channelencoding/decoding methods, frequency spreading/despreadingtechniques, and encryption/decryptionalgorithm.The various components of a software radio are shownin “Figure 4. and described below:Figure 2. Radio Signal Reception ProcessFigure 3. Radio Signal Transmission Process© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 117Figure 4. Block diagram of Software RadioA. Smart AntennaAntennas are conducting devices which transmit orreceive electromagnetic radiations. An antenna arrayconsists of distributed antenna elements whose outputsare combined and is a practical tool for enhancingwireless system performance. The choice of an antennafor a software radio is crucial as it is expected to supportmultiple bands [1].It is an antenna array system with in-built signalprocessing “smart” algorithms to help adapt to differentsignal environments. It can also identify the direction ofarrival of the signal; add phases to the signal to create aconstructive radiation pattern to nullify interference.Smart antennas mitigate fading through diversityreception and beam-forming while minimizinginterference through spatial filtering. Software radiosprovide the flexibility needed for effective smart antennasand smart antennas put help in effective implementationand utilization of software radio.B. RF Front EndIt appears before the Intermediate Frequency changestate and after the signal is received from the antenna. Itdoes the job of converting the incoming signal to IFfrequency by using a tunable local oscillator.RF front end design process for a software radio hasthe challenge of catering to a variety of waveforms withwidely changing parameters such as amplitude,frequency, phase etc. Due to this wide spectrum, thepresence of noise and interference is manifold. Thismakes the process of achieving a dynamic range evenmore difficult. The dynamic range is the measure of thehighest and lowest level signals that can besimultaneously contained in a radio. In case of mobilecommunication, increase in dynamic range would meanmore battery consumption, which becomes a major tradeofffor mobile phones.Low level signals suffer from the problem of noise.Noise enters at the bottom of dynamic range due to thethermal effects of the components, deviation inquantization or sampling aperture jitter in an ADC. Highlevel signals are limited by interference. Interference iscaused at the high end due to adjacent channel, cochannelor is self induced by transceiver. Traditionalwireless communication receivers require single RF frontend for each channel. Software radio, however, has onlyone wideband RF front end which can digitize desiredsignal into separated channels via software to provide alow cost solution.C. Analog to Digital ConversionBefore all the processing can be done in the software,the analog signal received at IF stage needs to beconverted to digital samples by use of Analog to Digitalconverter. Ideal Software Radio would require a widebandwidth and good dynamic range from the ADC. Asshown in “Figure 5. , the translation of the signal fromanalog to digital is performed via sampling andquantization. Sampling changes the signal that existscontinuously in time to a signal that is non zero only atdiscrete intervals of time. Quantization changes thecontinuous valued signal to discrete valued signal.Figure 5. Analog to Digital Conversion© 2010 ACADEMY PUBLISHER

118 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010In an ideal software radio, Analog to Digitalconversion should occur near the RF end. This wouldrequire the following criteria’s to be met:• A very high or variable sampling rate to supportwide signals bandwidth.• Greater number of quantization bits to support ahigh dynamic range.• Operating bandwidth of several GHz to allowconversion of a signal over a wide range offrequencies.• Lesser distortion and wide dynamic range to allowrecovery of low level signals in the presence ofstrong interferers.• Economical components to meet the abovecriteria’s without consuming an excessive amountof power.Due to the limitations of the fabrication technology,few of the above mentioned criteria’s need to be tradedoff for an acceptable design solution for the softwareradio.D. Digital to Analog ConversionIn Software Radio, entire waveform and modulation isgenerated in digital domain and is then fed to Digital toAnalog Converter before final transmission. As shown in“Figure 6. , the translation of the signal from digital toanalog is performed via voltage mapping andreconstruction, where the digital signal is changed to acontinuous valued signal. The number of bits and thefrequency range of the DAC are the important factors asthey would determine the reconstruction of thecontinuous signal.In a traditional analog radio, the frequency synthesis isperformed via bulky devices like quartz crystal, inductor,capacitor, mechanical resonators etc. But differentfrequencies and waveforms can be generated in thedigital domain by Direct Digital Synthesis (DDS)techniques and it is fast replacing the analog devices asthey provide better accuracy, frequency can be changedwith software and is simple to implement.E. Digital Down ConversionA digital down converter (DDC) provides the linkbetween the analog RF front end and the digital basebandof a receiver. The received signal is usually sampled atmuch higher sampling rates than required to relax thespecifications of the anti-aliasing filter required. But toreduce the computational power required from DigitalSignal Processors (DSPs), the sampling rate isimmediately reduced by down converter. DSPrequirements are directly proportional to sampling rateand reducing sampling rates reduces CPU cycles requiredto perform the same digital signal processing algorithmsand thus saving significantly on cost and DSP powerconsumption.F. Digital Up ConversionA digital up converter (DUC) provides the linkbetween the digital base band and analog RF front endand is required on the transmitter end. The signal to betransmitted is generated by digital signal processing atlower sampling rates to reduce computations of DSP. Butbefore it can be fed to DAC for converting to analogsignal, it is up converted to higher sampling rate byDigital Up Converter to relax interpolation filterspecifications.G. Digital Signal ProcessingDSP is the brain behind all digital radio technology. ADSP core consists of an arithmetic logic unit (ALU),accumulator(s), multiply and accumulate MAC unit(s),data and the address buses. A DSP is designed to supporthigh performance, repetitive, numerically intensive tasksand very high I/O performance. Large accumulators inDSPs help reduce the precision problems. The digitalsignal processor performs all the functions of filtering,demodulating, generating the signal for transmissionusing various modulation techniques in software radio.DSPs also perform functions of data compression,encryption and special functions like speech recognition,image enhancement, neural networks for artificialintelligence etc. The three main categories of digitalhardware are ASICs (Application Specific IntegratedCircuit), FPGAs (Field Programmable Gate Array) andDSPs. A DSP represents the most generalized type ofhardware that can be programmed repeatedly using highlevel programming language to perform variousfunctions. An ASIC is the most specialized piece ofhardware and can be used only in the specific applicationfor which it has been designed. ASICs are generally usedwhen DSP has insufficient processing power and functionto be performed is fixed. The set-up of ASICs isimplemented on fixed silicon, which optimizes the speedand power consumption of the circuit.Figure 6. Digital to Analog Conversion© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 119FPGAs help conserve silicon area since one chip canbe configured to perform more than one function hencethe configuration can be done at run time. Gate arraysoffer higher degree of parallelism in comparison to DSP.A hardware description language (HDL) is used fordesigning highly complex circuitry. FPGAs are moreflexible than ASICs but lesser than DSPs. FPGAs arebeing used because of their ability to perform high-speedparallel multiplication and accumulation functions andare especially well suited to handle algorithms like FiniteImpulse Response (FIR) filters (for decimation orinterpolation) and transforms like Fast FourierTransformations (FFT) and Discrete Cosine Transforms(DCT). Also they have the advantage of providingflexibility to integrate logic design with signalprocessing. These three hardware components constitutea design space which trades flexibility, processing speedand power compensation.In an ideal scenario, Digital signal processing (DSP)software should help to perform most of the radiofunctions at astonishing performance levels, enhance theperformance using digital filtering and drasticallyreducing noise and interferences, all of it with less bulkyequipments.IV. MODULATION AND DEMODULATIONTo understand the advantage and ease with whichmodulated signals can be generated and demodulationcan be done in digital domain using digital signalprocessing, an example of Amplitude Modulation andDemodulation is discussed below.A. Modulating the signalThe carrier signal as shown in “Figure 7. , can begenerated using either DDS technique or using sine serieslibrary function. The modulating signal will be the outputof microphone amplifier sampled at a particular rate byADC. Here for illustration purposes, carrier frequency of1000Hz and modulation signal of 20 Hz has beengenerated using the following equations with samplingrate of 10 KHz:for ( i = 0; i < 1000, i++){Carrier = Sin(2 * PI * i * Fc / SR);ModS = 1 + Sin(2 * PI * i * Fm / SR);}.In (1) above,SR is sampling Rate = 10KHz,Fc = 1000Hz (Carrier),Fm = 20Hz (Modulation signal).Amplitude Modulation involves changing theamplitude of carrier as per the amplitude of themodulating signal. Since we have the samples of carrierand modulating signal in digital domain, the amplitudemodulation in digital domain is simply multiplication ofthese samples. “Error! Reference source not found.”,shows the generation of modulated signal.ModSignal = Carrier * ModS.(2)B. Demodulating the received signalAt the receiving end when this signal is received, it isfull wave rectified which in digital domain is nothing buttaking the absolute value of all the samples. “Error!Reference source not found. shows the output of fullwave rectifier.RectSignal = abs(ModSignal).(1)(3)Figure 7. Generation of carrier and modulation signalsFigure 8. Amplitude modulated carrierFigure 9. Full wave rectified AM waveform© 2010 ACADEMY PUBLISHER

120 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Now the modulated signal is contained in the form ofenvelope of full wave rectified signal and can berecovered by applying a low pass filter to remove thecarrier frequency. “Error! Reference source notfound.shows the output of a first order Infinite ImpulseResponse (IIR) filter. The equation of filter used is:Yn = 0.997*Yn-1 + 0.003*Xn.In (4) above,Xn is new input sample (RectSignal),Yn-1 is previous output sample,Yn is new output sample (demodulated signal).Figure 10. Demodulated signal after low pass filterV. SOFTWARE RADIO AS A SOLUTION TO CONVENTIONALANALOG RADIOSA Software Radio in contrast to a traditional Radioruns on a generic hardware platform to perform all themodulation/demodulation, frequency selection, filteringetc in digital domain as also explained above in theexample. It can perform role of a cordless phone, cellphone, internet access tool, a GPS receiver etc. all in one.This is possible by changing the software used forprocessing. Hence, there is no limitation of any particularfrequency range, modulation type, and output power in aSoftware Radio. All these flexibilities of software radiooffer advantages which can be broadly classified as:A. Frequency-AgileThe software programmable processing allows it tooperate in the entire radio waves frequency bands.B. Easy to plug n playIt is a radio into which multiple contractors could plugparts and software.C. Global MobilityIt caters to the model as it is flexible to support most ofthe communications standards like CDMA, GSM, IS-136etc.D. Inter-operabilityIt can be made to inter-operate with different wirelessprotocols, incorporate new services, and upgrade to newstandards etc by simply downloading a new program (or anewer version of the software) or by change of selectionfrom the user interface.E. Compact and Power EfficientDue to reduced number of components used inSoftware Radio, its size is reduced resulting in ease ofmanufacturing. Also less hardware results in less powerconsumption.(4)F. Quality of ServiceAs Software Radio is Software dependent; hence agreater quality of service is delivered consistently.VI. APPLICATIONS OF SOFTWARE RADIOThe ultimate goal of Software Radio is to provide asingle radio trans-receiver which can play the roles ofcordless telephone, cell phone, wireless fax, wireless e-mail system, pager, wireless videoconferencing unit,wireless Web browser, Global Positioning System (GPS)unit, and other functions still in the realm of sciencefiction, operable from any location on the surface of theearth, and perhaps in space as well. Few of theapplications of software Radio has been coveredhenceforth and shows the variety of functionality towhich it can cater to.A. Pacemakers and Implantable Cardiac DefibrillatorsWirelessly reprogrammable and implantable medicaldevices (IMDs) such as pacemakers, implantablecardioverter defibrillators (ICDs), neurostimulators, andimplantable drug pumps use embedded computers andradios to monitor chronic disorders and treat patients withautomatic therapies. For instance, an ICD that senses arapid heartbeat can administer an electrical shock alsoknown as the software radio attacks to restore a normalheart rhythm, then later reports this event to a health carepractitioner who uses a commercial device programmerwith wireless capabilities to extract data from the ICD ormodify its settings without surgery. Clinical trials haveshown that these devices significantly improve survivalrates in certain populations. [4]B. Smart sensorsSoftware Radios can be implemented as smart sensorsto determine the amount of various particles present inatmosphere, hence useful for pollution check, weatherforecast etc. When used as Bio-sensor with configurablesoftware, it can be used in study of complex anatomy.C. One Cell phone for many standardsIt is highly inconvenient for a regular inter countrytraveler to change his mobile connection forcommunication due to network issues. For example,3GPP UMTS is popular in European nations, whileC/TDMA standards are widely used across North andSouth America. So, every time one travels through acountry he has to have the country specific cellularphone. But with the coming of Software radio, it can be athing of past as the Software Radio can adapt to change inthe local air interface by a simple software change. Allthe base stations and terminals using Software Radioarchitecture can support multiple air interfaces duringperiods of transition and be easily upgraded.D. Cost Effective Up gradation of a Cellular BaseStationThe growth in radio technology is in-step with thedemand of today’s customer, but is the business ready?The customer may have asked for audio files in his© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 121cellular services yesterday, but today they look for videostreams in their phone and tomorrow they might ask forlive data feed. The cellular industry cannot afford to losebusiness by not making such changes. Software Radiohelps to save them millions by easy upgrades in software.There is no need for investment in the hardware, savingthem both money and time. It is also a boon forcustomers as they don’t need to purchase newer handsetsfor newer technology. Just as any individual or businesscan update a software program used on a PC, by usingSoftware Radio technology, any cellular providers canupgrade software easily and quickly to make changes ontheir systems. As a result, the cost for cellular phoneservice can continue to be decreased for consumers;while the providers can see higher margins for the servicethey provide.E. Usability in EmergencyAs the software radio is compact and easilydeployable, it can be useful for temporary coverageduring emergencies and special events. A software radiocan be programmed and mounted onto a vehicle anddriven to the site of a natural disaster to bring incommunications where they may have been damaged andto support multiple standards.F. High Definition TV ApplicationsSoftware Radios are finding significant applications inthe conversion of both analog and digital transmissions toHDTV, an upcoming broadcast technology.VII. CHALLENGES & LIMITATIONSSoftware radio technology alters traditional radios inthree ways namely moving A/D closer to antenna,substitute hardware with software processing andfacilitating transition from dedicated to general-purposehardware. This decreases the number of hardwarecomponents resulting in high degree of extensibility for awide range of features. This will also eliminate theredundant hardware as found in dual band phones etc.There are several technological challenges faced inSoftware Radio today, like the usage of faster A/Dconverters, requirement for less power consumption,implementing smart antennas and making the overallradio more cost effective. ADCs which can digitize thehigh range signal are very expensive. They also increasethe level of SFDR (spurious free dynamic range). Hence,the signal received are down converted from RF signaland digitized at IF range via ADCs before sending thedata to DSPs. To overcome this limitation, more researchand development would be required to build lessexpensive ADCs covering more dynamic range andsupporting greater sampling rates. For FPGAs, interval toreprogram FPGA’s limits its usability.The authenticity and security of downloaded softwareis another area of concern. There is a need for a globallyrecognized regulatory body which can control theseprocesses. Certification from bodies like FederalCommunications Commission (FCC) is another hurdle.FCC certification process presumes that any software thatis bundled with the hardware be certified as a unit. Butsoftware radio being programmable at any time poseschallenges in certification. A new certification procedurehas been created. Software changes which can affectradio frequency, power and modulation are subject tomore streamlined process. However, FCC does not certifyany changes by third party software vendors. Allmodifications are the responsibility of the originalmanufacturer whose original hardware/software radiobundle was certified. This requires that all theindependent vendors need to work via equipmentmanufacturers. Hence, certifying software radios pose asignificant challenge because of cumbersome certificationprocesses, especially for domains like aerospace anddefense which require testing for all possible scenarioswhere software may misbehave.VIII. FUTURE OF SOFTWARE RADIOUltimate aim of software radio would be to move ADCand DAC just next to Antenna and doing everything insoftware. Currently, the cost of high frequency ADCs andspecialized hardware for digital signal processing are fewof the factors limiting the application of software radiosto high end use only. As the technology advances, thecosts will come down and processing power of generalpurpose computing platforms will increase. Once theprocessing power crosses the threshold required forperforming necessary radio functions, broaderapplications and scales of economy will come into placeand we will see software radios being used in our day today use gadgets. How soon it can happen is still an openquestion but software radio is likely to be an importanttechnology in the years to come.REFERENCES[1] Jeffrey H. Reed, “Software Radio”.[2] J Mitola, "The Software Radio", IEEE NationalTelesystems Conference, 1992 - Digital Object Identifier10.1109/NTC.1992.267870.[3] Ulrich L. Rohde, "Digital HF Radio: A Sampling ofTechniques”, Ham Radio Magazine, April, 1985.[4] http://www.secure-medicine.org/icd-study/icd-study.pdf[5] “A Software-Defined Radio for the Masses, Part 1- 4”,Gerald Youngblood, AC5OG[6] http://www.sdrforum.org/[7] http://rfdesign.com/mag/radio_fast_hot_data/index.html[8] http://www.dspguide.com/© 2010 ACADEMY PUBLISHER

122 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010SDR and Error Correction using ConvolutionEncoding with Viterbi DecodingShriram K. VasudevanVIT/Embedded Systems, Vellore, IndiaEmail: shriramkv@rocketmail.comSiva Janakiraman and Subashri VasudevanSASTRA/Communication, Tanjore, IndiaSASTRA/Computer Science, Tanjore, IndiaEmail: sivajanakiraman@gmail.com,vrsuba@gmail.comAbstract— The aim of the paper is to build asoftware based radio capable of demodulating FrequencyModulated signals with the ability to receive upto fourstations simultaneously. In addition convolution encodingalong with Viterbi decoding is also implemented to aid inerror correction capability of digital transmissionsystems. The initial phase is aimed at software baseddemodulation of multiple channels of FrequencyModulated signals. The second phase is directed towardsthe Error correction using convolution encoding withViterbi decoding for streaming data encountered indigital broadcast systems.Software Radio is a way of designing softwarevery close to the antenna. The basic feature of softwareradio is that software will define the transmittedwaveforms, and software will demodulate the receivedwaveforms. This paper is developed with the help of freesoftware based radio toolkit given by a community namedGNU radio. Channel coding schemes need to be usedextensively in the case of transmission of digital data overthe air or through any other medium. These codingschemes also called FEC (Forward Error Correction)are used in cases where the re-transmission of the data isnot feasible or possible. The channel coding schemescomes to the rescue in such cases where redundant dataare sent over the transmission medium along with themessage. In the receiver side, this channel coded signal isdecoded to get back the original data even if the channelcoded signal undergoes some interference from the noisein the transmission medium. Though the channel codinghas a downside of requirement to transmit additionaldata over the transmission medium, the advantagesoffered by error detection and correction mechanisms tothe receiver outweighs the disadvantage of additionaldata transfer rate and bandwidth requirements.Index Terms—SDR, Viterbi Decoder, Convolution EncoderI. INTRODUCTIONSoftware-Defined Radio (SDR) is an evolvingtechnology that has received enormous recognition. In thepast few years, analog radio systems are replaced bydigital radio systems for various radio applications incivilian, military and commercial spaces. In addition tothis, programmable hardware modules are playing a vitalrole in digital radio systems at different functional levels.SDR totally aims to take benefits of these programmablehardware modules to get open-architecture based radiosystem software.SDR technology supports implementation of some ofthe functional modules in a radio system such asmodulation/demodulation, signal generation, coding andlink-layer protocols in software. This assists in gettingreconfigurable software radio systems where dynamicselection of parameters for each of the above-mentionedfunctional modules is possible. A Complete hardwarebased system has many limitations.Before getting into what software radio does, it is goodto review the design of a traditional analog, hardwarebasedradio (Figure.1). In wireless communications,information is encoded into radio waves. These arecollected (or transmitted) from (to) the air by the antenna.The received signal is then passed to a series ofcomponents that extract the useful information andconvert it into the output of the radio. The basic design isthe same whether the radio signal is destined for a cellphone, microwave repeater, or AM/FM car radio.Traditional radios are based on the super heterodyne(superhet) receiver circuit.Manuscript received August 2, 2009; revised August 20, 2009;© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.122-130

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 123key reasons why software radios will not bedeployed first in end-user devices suchas cell phones, butrather inbase stations which can take advantage ofexternal power sources.Figure 1. Generic FM receiverIn the superhet receiver, the incomingsignal, whichhis Radio Frequency (RF), is first down-converted to alower intermediate frequency(IF). The IF is then filteredfor noise andamplified before being demodulated toproduce the baseband signal that represents the desireddinformation. This analog baseband signal may be passeddirectly to further downstage processing, or it may firstbe digitized and subjected toadditional signal processing.There are several important reasons for down convertingto a lower and standardized IF:(a) It is easierand hence less expensive, to build filtersand amplifiers – especially linear amplifiers – for a lowerfrequency signal.(b) Use of a common IF enables standardization in thedesign of radiocomponents.Traditionaldesignsfor implementingthe superheterodyne receiver architecture were optimized forspecific frequencies and applications and each of thestages were implemented inhardware that was closelycoupled. This was due largely to the difficulties inherentin and limitations in the state of the art inthe design ofanalog signal processing components. Analog processingis much more complicated than digital processing, whichhis one of the key reasons why the transition to digitalsignals is so important.In this paper chapter 2 deals about theperformance and power limitations, Chapter 3 talks aboutarchitecture ofSDR. Chapter4 is on signalflow, Chapter5 is dealing with how to derive instantaneous frequency.Chapter 6 talksabout De Emphasizer. Chapter 7 explainsabout multi channel FM reception. Chapter 8 will givedetails about the Forward Error Correction.Chapter 9 and10 speaks on implementation and results. Chapter 11 isconcluding thepaper followed by references.II. PERFORMANCE ANDPOWER LIMITATIONSThe change from hardware to software radioshas faced lot of problems. First, performance generallycomes down in the shift from dedicated to general-purpose hardware.Secondly, the transition from hardware tosoftware processing results in a substantial increase incomputation which ultimately results in increased powerrequirements. This reduces battery life and is one of theIII. ARCHITECTUREWe now take a view at the overview of a basicconventional digital radio system and then look athowSDR technology can be used toimplement radiofunctions in software. This will be followed by thesoftware architecture of SDR. Thedefinition of thesoftware radio system is not absolute and depends onwhere theAnalog to digital conversion is performed. Thequestion of where theA/D conversion is performeddetermines what radiofunctions can be movedintosoftware and what types of hardware are required. At oneextreme, we consider calling it software radio if softwareis used atany stage within the radio.This case typicallyinvolves the digitization at the baseband for signalprocessing. In this casethe actual demodulation processis performed before theconversion todigital format. Thissetup is more of a digital radio than software radio. Whatprocess can be done using software isvery much limited.Since theactual demodulation itself takes placewithhardware, it cannot beused for demodulation of otherkinds ofsignal and hence is very limited in itsfunctionality.At the other extreme, we might choose to use softwareradio forcases in which A/D conversion is performedright at the antennaa with all radio functionalityimplemented in software running on general-purposehardware. In this case the Digitizationof the signaltakesplace right at the antenna and thus gives us the completeflexibility. This setup requires the use of a veryhighspeed andwideband ADC, not economically feasible atthis point in time. Figure 2 shows the basic softwareradio architecture.Figure.2 Basic Software Radio Architecturee© 2010 ACADEMY PUBLISHER

124 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010In this paper, we take a stance that lies inbetween the two cases described above. The digitizationof the signal is performed inthe Intermediate frequencyband, followed by the digital down conversion. Thebaseband processing is done in a general purposecomputer. Though this setupdoes not giveus a completeesoftware control over the radio spectrum, it does offer usa great level of flexibility with the software since we canactually demodulate the signal using software. Thearchitecture used in this project is indicated in Figure 3.This helps us to explore the part of radio spectrum whichhfalls in the IF region that canbe digitized.The digital radio systemconsists of three mainfunctional blocks: RF section, IF section and basebandsection. The RF section consists of essentially analoghardware modules while IF and baseband sectionsscontain digitalhardware modules. The RFsection (alsocalledas RF front-end) is responsiblefortransmitting/reeceiving the radio frequency (RF) signalfrom the antenna via a coupler and converting the RFsignal to an intermediate frequency (IF) signal. The RFfront-end on the receive path performs RFamplificationand analog down conversion from RF to IF. On thetransmit path, RF front-end performs analog upconversion andRF power amplification.The ADC/ /DAC blocks perform analog-to-digitaconversion (on receive path) and digital-to analogconversion (ontransmit path), respectively. DDC/DUCCblocks performdigital-downconversion (on receive path)anddigital-up-conversionrespectively. DUC/DDC blocks essentially performmodem operations, i.e., modulation of the signal ontransmit path and demodulation of the signal on receiveepath.For this paper work, the RF section andthe IF sectionare handled bythe hardware, followed bythe basebandprocessing by Computer. The hardware used is known asthe Universal Software RadioPeripheral (USRP).(on transmitpath) ),III. SOFFTWARE BASEDFM RECEIVERRThe real implementation of the project starts fromhere. The hardware used is a custom built board provideddby the creators of the GNU Radio community, which isdesigned to run the free software toolkit provided bythem. The main goal isto take the FM baseddemodulationone step further to receive up to four FMstations simultaneously.What theUSRP gives to the Computer is a Digital-Down converted, complex, quadrature signal in theBaseband. Theremaining processing willl be taken careby the software after gettingg the signal from the USRPboard. Thus once the signal enters the computer, thedemodulation of the FMsignal consists of the followingsteps.(a) Signal flow from the air to the computer (from real tocomplex)(b) Getting the instantaneous frequency (from complex toreal)(c) De-emphasizer(d) AudioFIR decimation filter(e) Outputo Soundcard/FileThus each stage of this signal processing acts as a blockwith input and outputt ports. Each block receives thesignal asinput from the previous block, performs therequired signal processing and gives the output to thenext block in the chain.The software architecture is implemented in two layers.The bottom layer is implemented in C++, which satisfiesthe performance requirement of thedemodulationandsignal processing. The top layer is implemented inpython, which acts as a mask layer, interconnecting thebottom C++ layers which perform the bulk of the signalprocessing. This two layered architecture gives us theadvantageor reusing the signal processing blocks andmakes it easier to customize the signal processingflownecessary. Thus connecting or removing a processingblock from the software demodulation chain is mucheasier. The entire software layer is built and run on a freeUNIX port for windows called cygwin. Cygwin offers aUNIX platform on windows based systemIV. SIGNAL FLOW FROM THE AIR TO THE COMPUTER(FROM REAL TO COMPLEEX).Basically what the USRP does is to select the part ofthe spectrum we are interested in and decimate the digitalsequence by some factor N. The resulting signal iscomplex with I/Q two channels. Thus USRP gives out a`complex' signal, witha data rate 256k samples persecond, called `quadrature rate' – or quad rate becausethe complex signal has I/Q quadrature components(Figure. 3).Figure.3 Digital down Converter© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 125V. GETTING THE INSTANTANEOUS FREQUENCY (FROMCOMPLEX TO REAL).For FM, the instantaneousfrequency of thetransmitted waveform is varied as a function of the inputsignal. The instantaneous frequency at anytime is givenby the following formula:f(t) = k * m(t) + f cm(t) is the input signal, k isa constant that controls thefrequency sensitivity and f c is the frequency of the carrier(for example, 100.1MHz). So to recover m(t), two stepssare needed. First we need toremove the carrier f c , thenwe're left with a baseband signal that has aninstantaneousfrequency proportional tothe originalmessage m(t) ). The secondstep is to compute theinstantaneous frequency of the baseband signal.Removing the carrier is taken care by the USRP, viathe digital down converter (DDC). The resulting signalcoming into the Computer has already become abaseband signal and the remaining task is to calculate itsinstantaneousfrequency. If we integrate frequency, weget phase, or angle. Conversely, differentiating phasewith respect totime gives frequency. These are the keyinsights we use to build the receiver. The angle betweentwosubsequentsamples can be determined bymultiplying one by the complex conjugate of the otherand then takingthe arc tangent of the product.Thus Arctangent of theproduct gives the phasedifference between adjacentsamples, if we divide it bythe sample interval or multiply the data rate, we get theradian frequency, which gives the instantaneoussfrequency (f) if further divided by 2. This process ofrecoveringtheinstantaneousfrequency from thequadrature signal from the USRP is called the“Quadrature Demodulation”(Figure. 4).Figure.4 Quadrature Demodulationfrequency end usually becomes unbearable. Tocircumvent this effect, `pre-emphasis' and `de-emphasis'were introduced into the FM system. At the transmitter,proper pre-emphasis circuits are used to manuallyamplify the high frequency components, and the converseoperations are done at the receiver to recover the originalpower distribution of the signal. As a result, the SNR isimprovedeffectively. Inthe analog world, a simple firstorder RLC circuit usually suffices for pre-emphasis andde-emphasis. In the case of the Software defined Radio, afirst order IIR filter is implemented to get the magnitudeof amplified, high frequency signal down tothemagnitude of normal lower frequency signal, therebyimproving the SNR.Let us now see about Audio FIR decimation filter:After passing the de-emphasizer, of 256kHz.It is a baseband signal, containingall the frequency components of a FM station. Theitis a real signal witha data ratebandwidth of a FM station is usually around 2 * 100kHz.A sample rate of 256kHz is suitable for the 200kHzbandwidth, without losing any spectrum informationnFigure.5 Typical FM BandThe FMsignal band is spread out with various band ofinformation (Figure. 5) ). From 0 to about 16 kHz is theleft plus right (L + R) audio. The peak at 19 kHz is thestereo pilot tone. The left minus right (L - R) stereoinformation is centeredat 2x the pilot (38kHz) and isAM-modulated on top of the FM. Additional sub carriersare sometimes found in the region of 57kHz - 96kHz.Thus for the reception of a mono FMstation, a LowpassFIR filterwith a pass band frequency of 15Khz and atransitionn band with 1Khz should suffice. Thus, once thesignal comes out of theFIR decimation filter block, it isnow ready to be played into the soundcard.Output toSoundcard/File:VI. DE-EMPLASIZERIt has beentheoreticallyproved that, in FM detector,the power of the output noise increases with thefrequency quadratically. However, for most practicalsignals, such as human voiceand music, the power of thesignal decreases significantlyas frequencyincreases. Asa result, the “Signal to Noise Ratio” (SNR) at the highOnce the signal passes through the FIR Decimationfilter, it is ready to be played bythe soundcard oralternatively stored in a File. The signal from theFIRdecimation filter is ina raw bit format and can beconvertedto other formats using any suitablecompression algorithms.© 2010 ACADEMY PUBLISHER

126 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010VII. FORWARD ERROR CORRECTIONForward error correction [FEC] is a system of errorcontrol for data transmission, whereby the sender addsredundant data to its messages, which allows the receiverto detect and correct errors (within some bound) withoutthe need to ask the sender for additional data. Theadvantage of forward error correction is thatretransmission of data can often be avoided, at the cost ofhigher bandwidth requirements on average, and istherefore applied in situations where retransmissions arerelatively costly or impossible.Such a typical situation where the receiver is not ableto make a request for the re-transmission of message isFM reception. FEC devices are often located close to thereceiver of an analog signal, in the first stage of digitalprocessing after a signal has been received. That is, FECcircuits are often an integral part of the analog-to-digitalconversion process.Forward Error Correction Implemented:The Forward error correction algorithm consists oftwo parts namely Encoding and Decoding. The Encodingis implemented using a Convolution algorithm and theDecoding implemented is the Viterbi algorithm. Theconvolution encoder and the corresponding Viterbidecoder are implemented in C programming language.Motivation for Forward Error CorrectionThe situation in which the receiver is not able to makea request to the sender for re-transmission of messageforms the basis for all forward error correctiontechniques. A typical radio receiver is a classic examplein which the receiver has no provision to communicatewith the sender. Under such circumstances, if an erroroccurred in the signal transmitted, there will be no wayfor the receiver to detect the error unless some errorcontrol mechanisms are incorporated into the signalbefore transmission itself. This process of adding errorcontrol to the signal before transmission is calledForward Error Correction.A typical communication channel in which the datato be transmitted is encoded using some encodingalgorithm by adding some redundant bits (Figure. 6).Thus the output of the encoder is called ChannelSymbols. These channel symbols are then transmittedover the wireless channel.Thus it can be seen that unless the error correctionmechanisms are introduced there is no way for thereceiver to recover the original information correctly.Hence Forward Error Correction mechanisms areessential if the receiver is to recover the originalinformation. There are numerous forward error correctionmechanisms each applicable to particular type ofcommunication. The factors which govern which errorcorrection algorithm is used for a particularcommunication type are Bandwidth of the spectrumDepth of error correction requiredRate of encodingProcessing capacity of Sender/ReceiverIntroduction to Convolution EncodingConvolution Encoding typically involves encoding ofstream of data bits by using previous bits to perform alogical operation, the output of which is the encodedchannel symbols. Thus depending upon the type ofconvolution encoding, a single data bit can have itsinfluence on two or more adjacent bits thereby providingthe required redundancy to the data. This influence whicha bit has on other bits is what enables the receiver toidentify any bit errors that occurred during the datatransmission.A convolution encoder is called so because it performsa convolution of the input stream with encoder's impulseresponses. It can be represented mathematically aswhere x is an input sequence, y j is a sequence from outputj and h j is an impulse response for output j. A convolutionencoder is a discrete linear time-invariant system. Everyoutput of an encoder can be described by its own transferfunction, which is closely related to a generatorpolynomial.Figure 6. Typical wireless communication channel© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 127C.1. Convolution Encoding Algorithm in this paperThe convolution algorithm implemented in this projectis a Half rate,(5,7) encoder with a constraint length ofK=3, which is suitable for streaming data. .Theconvolution encoding typically involves the process ofencoding data by using the bits of data to perform modulo2 addition, thereby adding redundancy (Figure. 7).Then the message with errors is given as input to theViterbi decoder and its error correction capability isanalyzed.Convolution EncoderThe length of the generated bit stream is taken as inputfrom the user. Once the bit stream is generated, theconvolution encoding is performed and the encoded datais written to a file. The encoder simulation is shown infigure.7Figure 6. Basic convolution encodingXIII. VITERBI ALGORITHMThe Viterbi algorithm implemented in this project issuitable for decoding data that has been encoded with aHalf-rate,(7,5) convolution encoder with a constraintlength of K=3.IX. IMPLEMENTATION OF ALGORITHMSThe steps involved in simulating a communicationchannel using convolutional encoding and Viterbidecoding are as follows1. Generate the data to be transmitted through thechannel.2. Convolutionally encode the data.3. Introduce channel errors4. Perform Viterbi decoding on the receivedchannel symbolsX. RESULTSThe results of the project implemented are analyzed inthe sections below.Software Based Demodulation:Forward Error CorrectionThe simulation setup consists of a channel encoderwhich generates random bit streams consisting of 0’s and1’s which represents the message bits. This bit stream isthen encoded. Then the encoded data is given to achannel error simulator to simulate bit errors.Figure.7 Simulation result of encoderSince the encoding is typically performed on a streamof data, it must have a known length for the decoder todecode the message properly. Hence the encoder encodesthe data in frames. The frame length used in this projectis 15 message bits with 3 flushing bits for each frame.Hence the stream of data is broken up into frames ofconstant length and then encoded. Once the encodingprocess is completed, the encoded message is then writtento a file.Channel Error SimulationIn order to verify the error correction capability of thedecoder, the encoded message must be included withrandom bit errors to simulate the message transmittedthrough a noisy channel. A model in which random biterrors occurs is created.The number of bit errors to be introduced in eachframe is taken as an input from the user andcorrespondingly random bits are inverted i.e. 0’s changedto 1’s and 1’s changed to 0’s which provides a reasonablemodel to simulate channel errors. The channel errorsimulation module is shown in figure 8© 2010 ACADEMY PUBLISHER

128 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010The Test case taken up consists of a 60-bit messageencoded into four frames by the encoder. This setup isshown in FigureFigure.8 Channel error simulationIn this case, the number of bit errors introduced is zeroand hence the encoder output and the received channelsymbols are identical. The output of the channel errorsimulator is written to a file, which will then be taken asthe input for the decoder.Decoding received symbolsThe input to the Viterbi decoder is the file containingchannel errors, given as output by the channel errorsimulator. The decoder takes this file as input and thenperforms decoding and tries to recover the originalmessage. The decoding simulation is shown in figure 9Figure 10. An example depictionThe message is encoded into four frames and ispassed to the channel error simulator to introduce variousamounts of bit errors and the decoder performance isanalyzed.Figure.9 Decoding simulation resultWhile encoding messages are encoded in frame, so thatdecoder can start from a known state for each frame.Figure 9 shows decoding of channel symbols containingonly one frame.Performance of Viterbi DecoderThe Viterbi decoder implemented in this project is aHard-decision decoder capable of decoding the channelsymbols encoded by half-rate,(7,5) convolution encoderwith a constraint length of 3. Various amounts of biterrors are simulated and the performance of the decoderis then analyzed.In the first case, three random bit errors areintroduced in each frame. The performance of thedecoder for this test case is shown if following ScreenShot.© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 129It is found that the Viterbi decoder is able to recover theoriginal message in most of the cases when 5 random biterrors are introduced in each frame, thus offering almost100% error correction when there is 33% error in thereceived channel symbols.The next test case involves introducing seven bit errorsper frame and the performance of the decoder in this caseis shown in following snap.It is found that the decoder recovers the originalmessage when three random bit errors are produced ineach frame of the received channel symbols, offering100% error correction capability for 3 random bit errors.In the next test case, Five bit errors are introduced in eachframe as shown in screen shot as follows.It is found that the performance of the decoderdeteriorates as the bit error increases to 7 errors perframe. The decoded messages differ from originalmessage and on an average only 32% of the errors arecorrected on a trial run with 10 samples.When the number of bit errors per frame is furtherincreased to 9 bit errors per frame, the performance of thedecoder still goes down and the original message in notrecovered by the decoder as indicated in next screen shot.The performance of the decoder for this test case isshown in the following screen shotIt is found that in case of nine bit errors per frame, theerror recovery rate falls to as low as 8.2% on average runwith 10 samples.© 2010 ACADEMY PUBLISHER

130 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Thus overall, the performance of the Hard-decisionViterbi decoder falls with the increase in bit error ratesand good level of error correction is available up to 5 biterrors per frame above which the message keeps gettingdeteriorated. This is indicated in the figure .11%Bit error correction120100806040200Bit Error Correction Rate1 2 3 4 5 6 7 8 9No of Bit errors/frameFigure 11. Analysis graphXI. CONCLUSIONThus the software based multi channel FM receptionoffers the capability to receive from one up to four FMstations at the same time. One station is streamed live tothe speakers while the rest are stored in the disk in rawformat. This offers the ability to listen to the recordeddata at a later time thereby preventing the user frommissing out on some important information.Working with the FM band of the radio spectrum is theentrance to the Software based radio world which iscapable of handling many parts of the radio spectrum.This project provided a good practical exposure to theproblems involved in replacing a conventional radio withsoftware based one and demonstrated the power andflexibility offered by using software in place ofconventional hardware.This project just touches the tip of the iceberg in theworld of software defined radio, which has tremendouspotential to be unleashed. With the aid of technologyadvancement, the software based radios will be able toexplore the full radio spectrum very soon into the future.With the move towards processing signal systems indigital format comes the problem of error detection andcorrection. Thankfully, this has been taken care bynumerous error detection and correction algorithms,which were proposed long back, but were notimplemented due to large computation requirements.The convolution encoder and the Viterbi decoderimplemented in this project offer a reasonable amount oferror correction capability to a message signal.These algorithms are the basic error correctionalgorithms on which many other complex error correctionalgorithms are built upon.REFERENCES[1] John G Proakis, Dimitris G Manolakis, “DigitalSignal Processing: Principles, Algorithms andApplications”, 1996, Prentice Hall, ISBN: 0-13-373762-4[2] Todd K. Moon, “Error correction coding:Mathematical methods and algorithms”,2002, JohnWiley and sons, ISBN: 0-471-64800-0Shriram K. Vasudevan was born at TamilNadu on 29-07-1983. He holds an M.Tech in Embedded Systems and B.E., inElectronics and Instrumentation. He has done his B.E., fromAnnamalai University, Chidambaram, India. He proceeded withhis M.Tech at TIFAC-CORE of SASTRA University, Tanjore,India.He has got overall 4 years of industrial experience in thefield of Embedded Systems and Optical Networking. Currentlyhe is holding the responsibility of Assistant Professor in thefield of Embedded Systems in VIT University, Vellore, India.He was employed with Wipro Technologies for 2 years. He hasvisited USA, Dallas for his telecom training.He has presented papers in National and Internationalconferences in the areas of VLSI, Embedded and Networking.His research area focuses on Embedded Networking.Siva Janakiraman was born at TamilNadu on 05-06-1982.He holds an M.Tech in Embedded Systems and B.E., inElectronics and Communication. He has done his B.E., fromSASTRA University, Tanjore, India. He proceeded with hisM.Tech at SASTRA University, Tanjore, India. He has gotoverall 5 years of teaching experience in the field of EmbeddedSystems. Currently he is holding the responsibility of AssistantProfessor in the school of Electrical and ElectronicsEngineering in SASTRA University, Tanjore, India. Hisresearch area focuses on Embedded Systems and Cryptography.Subashri Vasudevan, born at TamilNadu on 21-01-1989.She is currently doing her B.Tech and she is in her final year ofher studies. She is graduating at SASTRA University. She hasbeen consistently working on many projects and she is one ofthe few toppers in her department. She is currently doing projecton Software Radio.She has presented papers in the area of Software Radio andWireless communications. Her research interest includesWireless communications and Embedded Systems.ACKNOWLEGDEMENTWe wish to thank all the co authors who supported thisproject for its success.© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 131Webinar – Education through DigitalCollaborationAnuradha VermaInfosys Technologies Ltd, Chandigarh, IndiaEmail: anuradha_thakur@infosys.comAnoop SinghInfosys Technologies Ltd, Bangalore, IndiaEmail: anoop_singh@infosys.comAbstract—Transition in the learning habits and trends ofstudents is evident from the traditional text based to thecurrent dynamic modes like world-wide web, CAI andsimulation. Webinar as a learning technology offers aplatform to overcome the gap in this digital divide and helpthe students get acquainted with the latest. A case describingthe use of webinar by Infosys helps the reader comprehendthe digital collaboration webinar offers discussing itsshortcomings and benefits.Index Terms—webinar; learning habits, teachingmethodologyI. LEARNING TECHNOLOGIESAs per Gretchen Rhines Cheney et al. (2005) [1] Indiahas the second largest education system in the world,falling only behind China. University EducationCommission (UGC) [2] set up in 1948 provides guidelinesfor coordination, determination and maintenance ofstandards of university education in India. Studying thevast historical background of education in India, it can beseen that though the Indian education system is powerful,yet it needs to dynamically revamp various facets toemerge strong in today’s ever changing scenario.In today’s fast pace world students are well informedand require a facilitator rather than an instructor. Theyhave islands of information available but do not have thepatience to understand its use or consult their faculty. Thelearning technologies have also seen a change in thelearning habits of students from the past to the recenttimes. The current day students are well versed with themodern technologies and are avid internet userscompared to their predecessors who were morecomfortable in a text based instruction based approach.However in the current education system, there exists adigital divide – a gap in the content and method ofknowledge dissipation and what value it offers to itsstudents in today’s ever changing world of technologycollaboration. Studying the trends in the past, a shift canbe ascertained in learning habits of students from thetraditional method of lecture based to the interactive anddescriptive based on World Wide Web (WWW) orinternet, e-mentoring, etc.This paper talks about the shift in the learning habitsand technologies and the impact webinar can play as alearning technology taking a case study from CampusConnect team of Infosys Technologies Ltd.II. CHANGE IN LEARNING HABITS FROM TRADITIONALTO TECHNOLOGY BASEDTill late, the teaching methodology used in most of theeducational institutes in India was following thetraditional method of chalk and talk but in recent timesthis has seen a paradigm shift.A. Traditional MethodologyKey features of the traditional teaching methodologycan be collated as under:• One way communication without any interaction• No openness between the faculty and students asa forum to exchange ideas• Less focus on analytical skills and more onmemory based.• Only lecture based class room approachfollowed for knowledge deliveryThis approach faced a lot of challenges, some being:• Lack of Flexibility – the students were notoffered any flexibility in terms of their learningstyles and habits or of implementing theirlearning. There was a fixed approach beingfollowed.• Less or no access to Information - Improvingaccess of information to the students is difficultas there is was no mode for information sharing.• No Standardization – Quality check at all levelsof education was absent offering nostandardisation.B. Scientific MethodologyThe advent of multi-media revolutionised trainingtechniques and brought in greater diversity andinteraction. The black boards have been replaced by LCDscreens. Some of the ingenious methods using scientificor technology base are:© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.131-136

132 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010already carved a niche for itself in the arena of businessand has now started being use in education arena as well.After increasing the dynamism in the industry, webinarsare all set to bring a revolution in the Education sector.The focal feature of a webinar is its potential to discussand share information. Offering a one stop shop forinteracting with an array of experts, this platformprovides a great appeal to its users.Some more characteristics of webinars are discussedbelow:A. CharacteristicsFigure 1. Various Knowledge Dissipation Methods• Multi-media Aids – teaching using medium likepower point presentations and slides, audiovisualclipping, etc enhancing the learningexperience.• Computer aided instruction (CAI) - use ofcomputers in delivering training, supervisingtrainee progress, feedback and assessing results.A similar technique is Computer based trainings(CBT).• E-mentoring and E-learning – as thesetechniques are not dependent on the physicalpresence of the instructor; they help students tostudy on their own comfort and speed.• Video Conferencing – It incorporates voice,image and data during long distancetransmission offering a real time and highlyproficient means.• Brainstorming & Case studies – is a groupcreativity technique designed to generate a largenumber of ideas for the solution of a problem.• Simulation – reproduction of the real timeenvironment for introducing students to realtime challenges.• Webinar – seminar or lecture over the internet[3] .This study attempts to establish the role of Webinars –an upcoming technology in the field of digitalcollaboration - in bridging the digital divide.III. WHAT IS A WEBINAR‘Webinar’ is a union of ‘web + seminar’ whichsimply means a seminar over the internet. This softwareis a remarkable innovation in the field of technologywhich offers a platform for people to interact andcollaborate over vast geographical boundaries throughWWW. This platform offers a two-way communicationleading to higher effectiveness and involvement by theaudience.Typically a webinar consist of a presentation hosted bya service provider on a web server. The link of thewebinar is shared with the attendees who can log on tothe site and participate in it. The Webinar platform has• Sharing Application – It allows presenters toshare their desktops, applications, etc to help theaudience get a better understanding of the topic.• Chat window – Attendee students can ask theirqueries during the session without disturbing theflow of the session through chat window. Ithelps in interaction with the presenters privately,interaction with the panellists privately,interaction with other participants privately orinteraction with all the participants in one go.• Session Recording– It may be beneficial torecord the session delivered by the presenter forre-use. This is an out-of-box feature in webinars.The session can be recorded and shared with thestudents or participants in the form of CDs, etc.This also aids in archival of valuableinformation.• Survey – The presenter can choose to conductpolls and surveys for the audience.B. Infrastructure RequirementFor Organisers:• One dedicated personal computer (PC)• Online WEBINAR monitoring PC station(Preferred)• Dedicated phone system and line (high qualitypreferably Polycom)• Voice-Tap hardware for WEBINAR recordingFor Participants:• One dedicated personal computer (PC)• Installation of Software of the vendor being usedfor Webinar (For example WebEx [4] ,GoToWebinar [5] , etc)• Dedicated direct phone line• LCD projector• High quality speaker phone with amplifier (ifrequired for large audience) and mute buttonC. Steps to organise a Webinar© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 133Reaching-out DigitallyWebinar Library(Recorded)Subject & TopicsKey Questions to be Answered• Technical and Soft Skills• Faculty & ManagementSMEsPractitionersEducatorsBAsRegionEvent Center and Moderated AudioConference FacilityPodcast, Audio-Video Interviews, Internet RadioFigure 2. Webinar Event flowThe above diagram describes the event flow for awebinar. The host can send the invitation of the webinarevent to intended audience (Student, faculty, etc) throughan email, or post it on organization website or bulletinboard sharing the link in the invitation mail. On the‘Landing Page’ the logged in users can find detailedinformation about the webinar like objective, speakerprofiles, reading materials links, etc. For attending theevent, participants need to register on ‘Registration Page’.Once the registration is complete, participants receivean email detailing the next set of instructions for joiningthe webinar. A reminder mail can also be set for theevent. For the audio of the webinar event, participantsneed to use their telephone and a computer for viewingthe presentation. After the event the attendee details canbe used for collating feedback and follow-up.IV. IMPACT OF WEBINAR ON EDUCATION – ACASE STUDY OF INFOSYS CAMPUSCONNECT INITIATVECampus Connect Initiative was launched in 2004 ByInfosys Technologies Ltd as an industry- academiapartnership which with the aim of enhancing the qualityand quantity of talent pool in India.Campus Connect team had a target to reach 500+partner colleges in an proficient and cost effective wayfor sharing technical, soft-skills, and domain knowledge.This required reaching the colleges across tier 2, 3 citiesand deliver sessions which required approximately 5hours of travel. Having Infosys Subject Matter Experts(SMEs) from various units spend such a large amount oftime on travel wasn’t a viable approach.Hence, the team decided to use a technology platformcalled Webinar. To augment the planning andimplementation of webinars a process was put in place.The team decided to work with WebEx which is aglobally acclaimed online meeting applications andsoftware services provider. Once this was zeroed in, thenext challenge was to successfully deploy it across all 9Development Centres (DCs) of Infosys in India. It alsorequired each DC be equipped with the knowledge ofWebinar – understanding the concept, its set up andconduction.Figure 3. Webinar PlatformTo provide clarification for all of this and to define aformal process, the team came up with a Field Guide [6]for webinars which served as a single point of referencefor the above.The following give an idea on the effectiveness and useof webinar by the team. [7]A. Stakeholder Feedback1. Faculty feedbackThe webinar efficacy was calculated as part of theCampus Connect Perception Measurement survey in 306colleges with 521 faculty member responses. The averagerating turned out to be 3.53 on a scale of 1 to 4 (Scale: 1.Strongly disagree 2. Disagree 3. Agree 4. Strongly agree).2. Student feedback –Comprehensive feedback was collected for eachwebinar from the students. It covered feedback forcontent, flow, effective delivery, interaction throughQ&A sessions, etc. Average feedback scores for majorcategories are given in the table below. (Scale: 1–Poor, 2-Satisfactory, 3- Good, 4-Very good, 5-Excellent). OverallWebinar Effectiveness was rated as 4.29 out of 5, above'Very Good'.TABLE I.CONSOLIDATED WEBINAR FEEDBACKArea of FeedbackWebinar Content 3.93Webinar Design 3.90Presenter Effectiveness 4.05Planning & Delivery 3.95Overall Webinar Effectiveness 4.29B. Webinar MetricsAverage Rating (ona scale of 1-5)Some consolidated metrics for the webinars conductedso far across the DCs are mentioned below. Table 2 givesthe number of webinars conducted till date and theirimpact.© 2010 ACADEMY PUBLISHER

134 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010TABLE II.WEBINAR COVERAGE DATA PERIOD - FEB 2006 - MAY 2009, 9 DCSNo. Item Number1 Total number of Webinars conducted 592Number of Colleges impacted acrossIndia 5093 Number of Faculty impacted 11474 Number of Students impacted 22527C. Rationalization of Travel Cost and ProductivityTotal cost of 1 hour webinar session covering 10partner colleges is approximately Rs. 12,000/-(TwelveThousand) and to conduct similar seminar session in 1partner college by SME through physical travel cost Rs.39500 (Thirty Nine Thousand). So each webinar sessionresults in saving of Rs. 383000/- (Three Lack EightyThree Thousand).No.12TABLE III.COST BREAKUP TO CONDUCT 1 SEMINAR SESSIONCost Saving Per WebinarSessionRs.Cost of Webinar covering10 colleges 12000Cost of Seminar covering10 colleges395000 ( 39500 * 10Colleges)Total Savings per webinarsession 383000TABLE IVCOST SAVING PER WEBINAR SESSIONNo. Cost Breakup to conduct 1 Seminar Rs.1 Flight Tickets / Taxi Charges 120002 Accommodation – 2 Days 50003 Food /Misc. Expense 25004 Productivity loss due to travelling 100005 Banner / Marketing Material etc 10000Total 39500Savings in terms of the SME travel time is anothersignificant benefit. Taking a conservative estimate of twobusiness days the SME would have spent in travelling tothe college and delivering the seminar, a saving of 1018business days has resulted because of using webinars.D. Capability improvement – Scalability and ReachBefore the advent of Webinars in campus connectknowledge dissipation, SMEs were required to travel tothe colleges posing a restraint on the number of collegesthat can be covered. The reach to the colleges before andafter introducing webinars gives a clear picture of theextent of change brought about. Table 5 shows thenumber of colleges the team was able to reach before andafter introducing webinars.Without Webinars, the coverage percentage wouldhave come down from 22.18% to 18.53%. Because of thewide reach of Webinars, the coverage percentage hasincreased from 22.18% to 57.22%, an increase of 35.04%in coverage of colleges. This coverage has furtherincreased to 40% covering almost each Campus ConnectPartner Colleges.No.1234TABLE VIMPROVEMENT IN COLLEGE COVERAGEPeriod# PartnerColleges#CollegesCovered%CoverageNov 2005 - Oct2006 (WithoutWebinars) 275 61 22.18Nov 2006 - Oct2007 (WithoutWebinars) 367 68 18.53Nov 2006 - Oct2007 (WithWebinars) 367Nov 2007 – TillDate (With210(142+68) 57.22Webinars) 520 509 97.8E. Reusability and ReproductionThese webinars covered a wide range of topics fromTechnical, Soft-Skills, Effective English, Quality andDomain areas. These webinar sessions delivered by SMEare recorded and re-use later. These webinar sessions arereproduced and edited in such a manner that they can beused later on by partner colleges. Recording of thesesessions are shared with participants in the form of CDsor by uploading the content on WebEx portal. This alsoaid in archival of Digital library.V. BENEFITS WEBINAR PLATFORM USAGE IN INFOSYSCAMPUS CONNECTSome benefits the team achieved were:• Collaboration – It offered the presenters to shareinformation, data, applications, desktops, etcalong with allowing participants and presentersto interact amongst themselves publically orprivately thus serving as a successfulcollaborative medium.• Multiple speakers – The team could employ theadvantage of multiple speakers for a singlesession without bearing the travelling cost foreach. It has been seen that a panel of speakerscreates more impact than an individual,especially for longer presentations.© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 135• Greater target access – Webinar has an edge interms of its greater reach to audience andovercome geographical boundaries. It is easy toaccess and use through WWW and is availableanywhere anytime.• No location dependency – As webinar arelocation independent, it can involve SubjectMatter Experts (SME) from a variety oflocations to talk about their respective areas ofexpertise. This helped the students leverage theknowledge base of these SMEs which may nothave been possible otherwise.• Interactive platform – It offered an interactiveplatform to the session audience (students,faculty or college management) to communicatewith the facilitator hence making it an effectivemeans of communication. Having a questionanswer round with all the participants in a setorder like alphabetical helps solve the commonqueries of the participants and record them.• Learning offline – Recording allows participantsto re-run the session and learn the concepts indetail.• Engage senior speakers– as the presenters of thesessions were practitioners and seniorInfoscions, physical travel was not possiblegenerally. Webinars helped the team overcomethe issues and increase the SME base.• Download Material – Participants coulddownload the material of the webinar and read itoffline. This increased the learning window forthem. The SMEs could share applications duringtheir sessions as well.• Feedback –Instructor feedback could berecorded through webinars. It helped to analyseand improve further.VI. HURDLES FOR THE TEAM IN WEBINARS• Need for high end Infrastructure- requiresquality of service and has dependency ofelectricity connection especially for the VOIP(Voice over internet protocol). Call drop-offrates during webinar session is typically 10percent so live helpdesk support through phoneand online chat is required in case of anytechnical issue.• Cost of hosting – As a medium of instruction,webinar proves to be costly as compared to theothers. Institutions or students may find itdifficult to purchase licenses for webinarsoftware and sustain the same.• Restricted Audience Involvement–The two-waycommunication channel has a limited usagehere. Hence, this may prove to a hurdle at times.It may not be able to live up to the same level asa physical interaction with the expert.• Limited modes of multi-media usage –PowerPoint Presentation Animation along with videocannot be played smoothly if a participantdoesn’t have the required bandwidth. Webinardoesn’t allow the instructor to use audio andvideo mode simultaneously.VII.CONCLUSIONThe traditional methods of training in educationalsystem may have been in practice since a long time, thecurrent time throws a challenge for students to updatethemselves on the technologies and their usage. With theadvent of IT era, the learning habits of students haveundergone a transformation from referring the text booksto browsing the internet, having 3D images andanimations as aids.Webinar today has gained significant popularity. Moreand more institutions want to leverage technology in thefield of education and percolate its maximum advantageto their students. It has helped overcome the issues ofbandwidth and leverage expertise but has still to developits potential as an interactive forum and costeffectiveness.Going ahead, this technology can be deployed toharness a formal mechanism of measurement and anevaluation framework. Its usage can be extended to otherareas of learning and its target audience base increased.While the current study talks about the webinarsconducted through WebEx, some common vendors thatoffer this platform can be listed as below:• InterCall [8]• GoToMeeting [9]• IBM Lotus Sametime [10]• MeetingBridge [11]• ReadyTalk [12]• WebMeetLive [13]ACKNOWLEDGMENTWe would like to thank Dr Ramesh Babu, Mr. R.N.Prasad, Mr. Mukundh Nagarajan and Mr. Sarfaraz AbdulAziz Jaitapkar from Infosys Technologies Ltd. forsharing their inputs and experiences for this paper.REFERENCES[1] Gretchen Rhines Cheney, Betsy Brown Ruzzi and KarthicMuralidharan, “A Profile of Indian Education System,National Center on Education and the Economy,”November 2005[2] Home Page, http://www.ugc.ac.in/[3] Webinar Definition,http://www.pcmag.com/encyclopedia_term/0,2542,t=Webinar&i=54380,00.asp[4] Home Page, http://www.webex.com/[5] Home Page, http://www.gotowebinar.com/[6] Webinar Field Guide V1.0, Infosys Technologies Ltd.,2006[7] Anuradha Verma, Anoop Singh, “Leveraging Webinar forStudent Learning”, International Workshop on Technologyfor Education, 2009[8] http://www.intercall.com/© 2010 ACADEMY PUBLISHER

136 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010[9] https://www1.gotomeeting.com/?Portal=www.gotomeeting.com[10] http://www.ibm.com/lotus/sametime[11] http://www.meetingbridge.com/[12] http://www.readytalk.com/[13] http://www.webmeetlive.com/Anuradha Verma is currently working as TechnicalEvangelist in the Education & Research Department at InfosysTechnologies Ltd. She has over 6 years of IT Industryexperience covering areas like Training & Development,Testing, Software Development and Industry-AcademiaPartnership.She has Masters in Business Administration along with aMasters in Journalism and Mass Communication. She iscurrently pursuing her Doctor of Philosophy in Managementwith Human Resource Specialization. She has authoredpublications in International Conferences and Journal.Anoop Singh. At Infosys, Anoop Singh is currently workingas GROUP LEADER in the E&R Department. He has over 12years of IT Industry experience including 4 years of exposure inglobal market, working in various counties of South East Asiaregion like Singapore, Hong Kong and China. Anoop hasextensive IT Industry experience covering areas like Training &Development, Business Partner Development, Product Linemanagement and Industry-Academia Partnership.Anoop has earlier worked with IT organizations like NIIT,NIIT Shanghai Ltd. and have server large global customer likeMicrosoft, Singapore Airline, Beijing International Airport andvarious Chinese universities. He has worked on technologiesareas like Database, Network Design and various programminglanguages. Anoop also holds a Master’s Degree in InformationScience.© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 137Efficient Visual CryptographyEr. Supriya KingerCSE Department, Chitkara Institute of Engineering and Technology,Rajpura, Punjab, IndiaEmail: ahujasupriya@gmail.comAbstract – Visual cryptography scheme (VCS) is a secretsharingscheme which allows the encryption of a secretimage into n shares that are distributed to n participants.The beauty of such a scheme is that, the decryption of thesecret image requires neither the knowledge ofcryptography nor complex computation. Colour visualcryptography becomes an interesting research topic afterthe formal introduction of visual cryptography by Naor andShamir in 1995. It is a powerful technique which combinesthe notions of perfect ciphers and secret sharing incryptography with that of raster graphics. A binary imagecan be divided into shares which can be stacked together toapproximately recover the original image. Unfortunately, ithas not been used much primarily because the decryptionprocess entails a severe degradation in image quality interms of loss of resolution and contrast. Its usage is alsohampered by the lack of proper techniques for handlinggrayscale and color images. In this paper, I have developeda novel technique which enables visual cryptography ofcolor as well as grayscale images. The physical transparencystacking type of decryption allows for the recovery of thetraditional visual cryptography quality image. An enhancedstacking technique allows for the decryption into a halftonequality image. And finally, a computation based decryptionscheme makes the perfect recovery of the original imagepossible. Based on this basic scheme, I have then establisheda progressive mechanism to share color images at multipleresolutions. I extracted shares from each resolution layer toconstruct a hierarchical structure; the images of differentresolutions can then be restored by stacking the differentshared images together. I have implemented our techniqueand present results.Index Terms – Secret sharing, Color halftoning, imagesharing, multiple resolutions, secret sharing, and visualcryptographyI. INTRODUCTIONVisual cryptography was originally proposed for theproblem of secret sharing. Secret sharing is one of theearly problems to be considered in cryptography. Theidea of the visual cryptography model proposed in [1] isto split an image into two random shares (printed ontransparencies) which separately reveal no informationabout the original secret image other than the size of thesecret image. The image is composed of black and whitepixels. The original image can be recovered bysuperimposing the two shares. The underlying operationof this visual cryptography model is OR.Visual cryptography is a unique technique in thesense that the encrypted messages can be decrypteddirectly by the human visual system. Therefore, a systememploying visual cryptography can be used by anyonewithout any knowledge of cryptography. Anotherinteresting thing about visual cryptography is that it is aperfectly secure cipher. There is a simple analogy of theone time-pad cipher to visual cryptography.II. BACKGROUND ON VISUAL CRYPTOGRAPHYBesides introducing the new paradigm, Naor andShamir also provided their constructions of visualcryptographic solutions for the general k out of n secretsharing problem. One can assume that every secretmessage can be represented as an image, and furthermorethat the image is just a collection of black and whitepixels i.e. it is assumed to be a binary image. Eachoriginal pixel appears in n modified versions (calledshares) of the image, one for each transparency. Eachshare consists of m black and white sub-pixels. Eachshare of sub-pixels is printed on the transparency in closeproximity (to best aid the human perception, they aretypically arranged together to form a square with mselected as a square number). The resulting structure canbe described by a Boolean matrix M = (m i j ) n*m where m i j= 1 if and only if the j-th sub-pixel of the i-th share(transparency) is black. The important parameters of thescheme are:1. m, the number of pixels in a share. Thisparameter represents the loss in resolution fromthe original image to the recovered one.2. α, the relative difference in the weight betweenthe combined shares that come from a whitepixel and a black pixel in the original image.This parameter represents the loss in contrast.3. γ, the size of the collection of C 0 and C 1 . C 0refers to the sub-pixel patterns in the shares fora white pixel and black refers to the sub-pixelpatterns in the shares for the 1 pixel.The constructions can be clearly illustrated by a 2 outof 2 visual cryptographic scheme. Here we define thefollowing collections of 2*4 matrices:C 0 = all the matrices obtained by permuting thecolumns of1 1 0 01 1 0 0C 1 = all the matrices obtained by permuting thecolumns of1 1 0 00 0 1 1© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.137-141

138 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010The six patterns of shares created based on the abovematrices are shown in figure 1. Note that one pixel of theoriginal image now corresponds to four pixels in eachshare.(c) The second shareFig 1: The six patterns of 4 pixel shares: vertical,horizontal and diagonalA visual cryptography scheme can then beconstructed by picking shares in the following manner:a) If the pixel of the original binary image iswhite, randomly pick the same pattern 0 of f ourpixels for both shares. It is important to pick thepatterns randomly in order to make the patternrandom.b) If the pixel of the original image is black, pick acomplementary pair of patterns, i.e., the patternsfrom the same column in figure 1.It can be easily verified that the resultant scheme hasthe parameters [m = 4; α= 12 ; γ= 6]: any two shares ofC 0 cover two out f our of the pixels, while any pair ofshares from C 1 covers all the f our pixels. An example ofthe above scheme is shown in figure 2. The first image isthe original image, the next two are the shares and the lastimage is the recovered original image obtained byperforming the equivalent of physically stacking twoimage shares on top of each other (assuming that they areprinted on transparencies). It should be noted that the lastthree images in figure 2 are four times as large as the firstone but I have scaled them to the same size as the originalimage.(a) Sample of monochrome image(d) The stacked ImageFig 2. Implementation of existing methodologyIII. RELATED WORKThere has been a steadily growing interest in visualcryptography. Despite its appearance of being a simpletechnique, visual cryptography is a secure and effectivecryptographic scheme. Since the origin of this newparadigm, various extensions to the basic scheme havebeen developed to improve the contrast and the areas ofapplication have also been greatly expanded.In [1], the construction of (n,n)-VCS was extended for(k,n)-VCS. In 1996, the same authors introduced the ideaof cover based semi-group to further improve the contrast[3]. Ateniese et al. [4] provided the first construction of(2, n)-VCS having the best possible contrast for any n·2.Blundo et al. [5] provided a contrast optimal (3,n)-VCSand gave a proof on the upper bound on the contrast ofany (3,n)-VCS. [1] first considered the problem ofconcealing the existence of the secret image. [6] provideda general solution for that problem.The random nature of secret shares makes sharesunsuitable for transmission over an open channel. [6]used a modified scheme to embed some meaningfulimages into the shares. [7] used different moiré patternsto visualize the secret instead of different gray levels. Asfar as extending to color images goes, [8] provided aprimitive scheme for images of 24 colors. Hou [9] thenproposed a novel approach to share color images basedon halftoning. Other interesting topics include visualauthentication [10] and watermarking based on visualcryptography [11]. Recently, there has been an attempt tobuild a physical visual cryptographic system based onoptical interferometry [12]. However, all of these earlierworks result in a decrypted image of reduced quality.(b) The first Share© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 139IV. OUR CONTRIBUTIONThe state of the art in visual cryptography leads to thedegradation in the quality of the decoded images, whichmakes it unsuitable for digital media (image, video)sharing and protection. This is quite obvious in figure 2where the white background of the original imagebecomes gray in the decrypted image.Through this paper, I propose a visual cryptographicschemes that not only can support grayscale and colorimages, but also allow high quality images including thatof perfect (original) quality to be reconstructed.The nagging presence of the loss of contrast makestraditional visual cryptography scheme practical onlywhen quality is not an issue which is quite rare. I havetherefore focused our attention on specificallyovercoming this problem by primarily devoting ourefforts towards improving the quality of the reconstructedimages. I first extend the basic scheme from [1] to allowvisual cryptography to be directly applied on grayscaleand color images. Image halftoning is employed in orderto transform the original image from the grayscale/colorspace into the monochrome space which has proved to bequite effective.It is a well known fact that the digital halftoning isalways a lossy process [2], which means that whenever ahalftoning is used for the transformation, it is impossibleto fully reconstruct the original secret image. A newencoding scheme has therefore been developed whichallows for perfectly lossless transformation betweenmonochrome, grayscale and color spaces. This newencoding scheme can be seamlessly incorporated into theproposed scheme for visual cryptography and it allowsthe original secret image to be perfectly restored. Ibelieve this advancement in visual cryptography can beuseful in secret sharing of images, in transmission ofsecret images over multiple untrustworthy channels, in e-commerce of digital media and in digital rightsmanagement of digital media..(a) Sample of monochrome image(b) The first Share(c) The second shareV. OUR APPROACHA. For Monochrome ImagesUsing XOR to Fully Restore Monochrome SecretImages :-I, first made the crucial observation that withjust one additional computational operation; eventraditional visual cryptography can allow full recovery ofthe secret binary image. Normally, when we superimposethe two shares printed on transparencies, this stackingoperation is computationally modeled as the binary ORoperation which causes the contrast level to be lowered.By simply substituting this OR operation with the XORoperation, the original binary image can be recoveredwithout any loss in contrast. Thus, the produced imagecould have a more visually pleasant appearance with lessstorage space requirement. However, the XOR operationneeds computation - the physical stacking process canonly simulate the OR operation. Figure 3 recovers thesame secret image as in figure 2 using the XOR operationand thus it is clearly evident that the contrast of theoriginal image is restored.(d) The stacked Image with XORFig 3. Implementation of proposed methodologyAs we have seen earlier, the application of digitalhalftoning techniques results in some downgrading of theoriginal image quality due to its inherently lossy natureand it is not possible to recover the original image fromits halftone version. We will refer to this proposedscheme as EVCS (Efficient Visual CryptographicScheme).The novelty of my approach is that it not only allowsthe secret image to be just seen but allows the secretimage to be reconstructed with perfect quality. Theadvantage of this approach is that it still retains thecrucial advantages of traditional visual cryptography likesimplicity, visual decoding and perfect security. Theextra feature is that depending on whether additionalcomputing resources are provided, images of different© 2010 ACADEMY PUBLISHER

140 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010quality can be decoded from the same set of shares. Ifonly the stacking operation is allowed (i.e. nocomputations), then our scheme recovers the originalvisual cryptographic quality. If the XOR operation isprovided (instead of the OR operation of stacking), thenwe can fully restore the original quality image.B. For Colored Images (Halftone-based Grayscale andColor Visual Cryptography)Digital halftoning has been extensively used inprinting applications where it has been proved to be veryeffective. For visual cryptography, the use of digital halftoning is for the purpose of converting the grayscaleimage into a monochrome image. Once we have a binaryimage, then the original visual cryptography techniquecan be applied. However, the concomitant loss in qualityis unavoidable in this case.For color images, there are two alternatives forapplying digital halftoning. One is to split the color imageinto channels of cyan, magenta and yellow. Then eachchannel is treated as a grayscale image to whichhalftoning and visual cryptography are appliedindependently. After the monochrome shares aregenerated for each channel, channels are combinedseparately to create the color shares. This is the approachpresented in [9]. The alternative approach would be todirectly apply color halftoning, then perform theseparation into color channels followed by the applicationof visual cryptography to each channel independently.Actually, these two approaches lead to the same resultsfinally. There are many mature halftoning techniquesavailable for selection like dispersed-dot dithering,clustered-dot dithering and error diffusion techniques.Halftoning based visual cryptographic scheme can besummarized as follows:a) Encryption: This stage is for the creation ofshares. This can be further divided into the followingsteps:i. Color halftoning: Standard algorithms such as theones described in [2], [13] and [14] can be usedfor this step. One could do the color channelsplitting first and then do the grayscale halftoningfor each channel:ii.Or one could do color halftoning first followed bythe splitting:Creation of shares: Considering the case of(2,2)-VCS, the steps areb) Decryption: This stage is for the reconstructionof the original secret image. This can be further dividedinto the following steps:i. Stacking of shares: The following stacking (OR)operation needs to be performed:ii. Subsampling for reconstruction: Theseoperations need to be performed where everyblock of f our pixels is sub-sampled into onepixel of the final image. This step is optional andshould be used only with the XOR recoverydescribed in Section III-B.1 to achieve betterquality.Then, for every 2*2 block B(i; j) of I, whereIt is clear that our technique, though independentlydeveloped, is quite similar in spirit to the one described in[9]. So both share the same drawback that digitalhalftoning always leads to permanent loss of informationwhich means that the original image can never beperfectly restored. Inverse halftoning is a possiblesolution that can attempt to recover the image. The bestresults can obtain a restoration quality of 30 dB measuredin PSNR, which is quite good. But this is not sufficientfor applications which require that the original image befaithfully recovered. In fact, in all other cryptographictechniques, it is taken for granted that the decryption of aciphertext perfectly recovers the plaintext. But visualcryptography has been a glaring exception so far.© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 141VI. CONCLUSIONIn this paper, I have extended traditional visualcryptography by employing new schemes whichovercome its limitations. I propose a technique forgrayscale and color visual cryptography. Our insight isthat the OR operation in the traditional visualcryptography can be replaced by the XOR operation inorder to allow for lossless decryption. However, there aresome practical issues that need careful consideration.First, the transparencies should be precisely aligned inorder to obtain a clear reconstruction. Secondly, there isusually some unavoidable noise introduced during theprinting process. Thirdly, the stacking method can onlysimulate the OR operation which always leads to a loss incontrast. Proper alignment is absolutely essential whensuperimposing the shares. In real experiments, we havefound that obtaining perfect alignment is alwaystroublesome. As visual cryptographic schemes operate atthe pixel levels, each pixel on one share must be matchedcorrectly with the corresponding pixel on the other share.Superimposing the shares with even a slight shift inalignment results in a drastic degradation in the quality ofthe reconstructed image. In the worst case, even a singlepixel shift can render the secret image totally invisible.This alignment problem can be resolved if the boundaryof each share is clearly marked which can act as guidesfor the alignment.REFERENCES[1] M. Naor and A. Shamir, “Visual cryptography,” in Advancesin Cryptology -EUROCRYPT’94, A. D. Santis., Ed., vol.950. Springer-Verlag, 1995, pp. 1–12.[2] H. R. Kang, Digital Color Halftoning, ser. SPIE/IEE Serieson Imaging Science and Engineering, E. R. Dougherty,Ed.Bellingham, Washington USA and New York:Copublished by SPIE Optical Engineering Press and IEEEPress, 1999.[3] M. Naor and A. Shamir, “Visual cryptography 2: Improvingthe contrast via the cover base,” 1996, a preliminaryversionappears in “Security Protocols”, M. Lomas ed. Vol.1189 of Lecture Notes in Compute Science, Springer-Verlag, Berlin,pp.197-202, 1997.[4] A. D. S. G. Ateniese, C. Blundo and D. R. Stinson,“Constructions and bounds for visual cryptography,” in 23 rdInternational Colloquium on Automata, Languages andProgramming, ser. Lecture Notes in Computer Science, F.M. auf der Heide and B. Monien, Eds., vol. 1099. Berlin:Springer-Verlag, 1996, pp. 416–428.[5] C. Blundo, P. D’Arco, A. D. Snatis, and D. R. Stinson,“Contrast optimal threshold visual cryptography schemes,”SIAM Journal on Discrete Mathematics, available at:http://citeseer.nj.nec.com/blundo98contrast.html, vol. 16,no. 2, pp. 224–261, April 1998.[6] G. Ateniese, C. Blundo, A. D. Santis, and D. Stinson,“Extended schemes for visual cryptography,” TheoreticalComputerScience, vol. 250, pp. 143–161, 2001.[7] Y. Desmedt and T. V. Le, “Moire cryptography,” in the 7thACM Conference on Computer and CommunicationsSecurity’00, Athens, Greece, 2000.[8] V. Rijmen and B. Preneel, “Efficient color visual encryptionfor shared colors of benetton,” 1996, EUCRYPTO’96RumpSession. Availabe athttp://www.iacr/org/conferences/ec96/rump/preneel.ps.[9] Y. C. Hou, C. Y. Chang, and S. F. Tu, “Visual cryptographyfor color images based on halftone technology,”inInternational Conference on Information Systems,Analysis and Synthesis. World Multiconference onSystemics, Cyberneticsand Informatics. Image, Acoustic,Speech And Signal Processing: Part II, 2001.[10] M. Naor and B. Pinkas, “Visual authentication andidentification,” Lecture Notes in Computer Science, vol.1294, pp. 322–336, 1997. [Online]. Available:citeseer.nj.nec.com/67294.html[11] Q. B. Sun, P. R. Feng, and R. Deng, in InternationalConference on Information Technology: Coding andComputing (ITCC ’01), available at:http://dlib.computer.org/conferen/itcc/1062/pdf/10620065.pdf, Las Vegas, April 2001.[12] S.-S. Lee, J.-C. Na, S.-W. Sohn, C. Park, D.-H. Seo, andS.-J. Kim, “Visual cryptography based on an interferometricencryption technique,” ETRI Journal, vol. 24, pp. 373–380,2002, available at http://etrij.etri.re.kr/etrij/pdfdata/24-05-05.pdf.Supriya A. Kinger was born in Haryana,India on 15 th july 1982. She did herMaster’s in technology in field ofcomputer science from YMCA, Haryanain year 2005, India and Bachelor’s intechnology in Computer Science fromKU, Haryana, India. In year2003.Currently she is doing research infield of Software Engineering(Component Based SoftwareEngineering).She has more than 5 Years of experience in teaching andresearch. She has attended and organized a number ofworkshops and conferences. She hosted a conference ASET-2006 at CIET and acted as convener in it. She has presentednumber of papers in national and international conferences ofrepute on the topics of web crawlers, Component basedsoftware Engineering, Information Security and Networks.Currently she is Working with Chitkara Institute of Enginneringand Technology, Punjab, India as Senior Lecturer. Earlier shehas worked at Institute of Engineering and Technology. She islife member of ISTE© 2010 ACADEMY PUBLISHER

142 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Multistage Interconnection Networks: ATransition from Electronic to OpticalRinkle Rani AggarwalDepartment of Computer Science & Engineering,Thapar University, Patiala –147004 (India)raggarwal@thapar.eduDr. Lakhwinder Kaur, Dr. Himanshu AggarwalDepartment of Computer Engineering,Punjabi University, Patiala–147002 (India)mahal2k8@yahoo.com, himanshu@pbi.ac.inAbstract—Optical communication are necessary forachieving reliable, fast and flexible communication.Advances in electro-optic technologies have made opticalcommunication a reliable networking choice to meet theincreasing demands for high bandwidth and lowcommunication latency of high-performancecomputing/communication applications. So optical networksgives high performance as well as low latency. Althoughoptical MINs hold great promise and have advantages overtheir electronic networks, they also hold their ownchallenges. This paper compares electronic and OpticalMINs. The design issues and solution approaches availablefor optical MINs are also explained and analyzed.Index Terms— Multistage Interconnection Networks (MIN),Optical networks, Crosstalk, Window methods.I. INTRODUCTIONTo meet the increasing demands of high performancecomputing applications for high channel bandwidth andlow communication latency, traditional metal-basedcommunication technology used in parallel computingsystems is becoming a potential bottleneck. Now, theneed arise either for some significant progress in thetraditional interconnects or for some new interconnecttechnology be introduced in parallel computing systems.Electro-optic technologies have made opticalcommunication a promising network choice to meet theincreasing demands with its advancement in thetechnology. Fiber optic communications offer acombination of high bandwidth, low error probability andgigabit transmission capacity.Multistage interconnection networks have beenextensively accepted as an interconnecting scheme forparallel computing systems. As optical technologyadvances, there is considerable interest in using opticaltechnology to implement interconnection network andswitches. A multistage interconnection network iscomposed of several stages of switch elements by whichany input port can be connected to any output port in thenetwork. Optical MIN represents a very important classof interconnecting schemes used for constructing Opticalinterconnections for communication networks andmultiprocessor systems. This network consists of Ninputs, N outputs and n stages (n=log 2 N). Each stage hasN/2 switching elements each SE has two inputs and twooutputs connected in a certain pattern. The most widelyused MINs are the electronic MINs. In electronic MINs,electricity is used, since in optical MINs, light is used totransmit the messages. Although electronic MINs andoptical MINs have many similarities but there are somefundamental differences between them. Available opticalMINs were built mainly on banyan or its equivalent (e.g.baseline, omega) networks because they are fast in switchsetting (self-routing) and also have a small number ofswitches between an input-output pair. Banyan networkshave a unique path between an input-output pair, and thismakes them blocking networks. Non-blocking networkscan be constructed by either appending some extra stagesto the back of a regular banyan network. Crosstalk inoptical networks is one of the major shortcomings inoptical switching networks, and avoiding crosstalk is animportant for making optical communication properly. Toavoid a crosstalk, many approaches have been used suchas time domain and space domain approaches. Becausethe messages should be partitioned into several groups tosend to the network, some methods are used to findconflicts between the messages.II. MULTISTAGE INTERCONNECTIONNETWORKSMultistage interconnection networks (MINs) consistof more than one stage of small interconnection elementscalled switching elements and links interconnecting them.Multistage interconnection networks (MINs) are used inmultiprocessing systems to provide cost-effective, highbandwidthcommunication between processors and/ormemory modules. A MIN normally connects N inputs toN outputs and is referred as an N × N MIN [9,10]. Theparameter N is called the size of the network. There areseveral different multistage interconnection networks© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.142-147

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 143proposed and studied in the literature. Figure 1 illustratesa structure of multistage interconnection network, whichare representatives of a general class of networks. Thisfigure shows the connection between p inputs and boutputs, and connection between these is via number ofstages. Multistage interconnection network is actually acompromise between crossbar and shared bus networksof various types of multiprocessor interconnectionsnetworks [1]. Multistage interconnection networksattempt to reduce cost and decrease the path lengthspeed mismatch occur due to the high speed of opticalsignals. The second approach is all-optical switching.This has removed the problem that occurred with thehybrid approach but, such systems will not becomepractical in the future and hence only hybrid opticalMINs are considered. In hybrid optical MINs, theelectronically controlled optical switches, such as lithiumniobate directional couplers, can have switching speedsfrom hundreds of picoseconds to tens of nanoseconds.A. Switching in optical networksIn optical networks, circuit switching is used. Packetswitching is not possible with Optical MultistageInterconnection Networks. If packet switching is used,the address information in each packet must be decodedin order to determine the switch state. In a hybrid MIN, itmeans it require conversions from optical signals toelectronic ones, which could be very costly [4]. For thisreason, circuit switching is usually preferred in opticalMINs.Figure 1: A MultistageNetworkMINsTABLE IPROPERTIES OF DIFFERENT INTERCONNECTIONTECHNIQUESOPTICALELECTRONICProperty Bus Crossbar MultistageSpeed Low High HighCost Low High ModerateReliability Low High HighConfigurability High Low ModerateComplexity Low High ModerateHYBRIDGUIDEDWAVEALL OPTICSSPACEOPTICSIII. OPTICAL MULTISTAGE INTERCONNECTIONNETWORKSAn optical MIN can be implemented with either freespaceoptics or guided wave technology. It uses the TimeDivision Multiplexing. To exploit the huge opticalbandwidth of fiber, the Wavelength DivisionMultiplexing (WDM) technique can also be used. WithWDM, the optical spectrum is divided into manydifferent logical channels, and each channel correspondsto a unique wavelength. Optical switching, involves theswitching of optical signals, rather than electronic signalsas in conventional electronic systems. Two types ofguided wave optical switching systems can be used [5].The first is a hybrid approach in which optical signals areswitched, but the switches are electronically controlled.With this approach, the use of electronic control signalsmeans that the routing will be carried out electronically.As such, the speed of the electronic switch control signalscan be much less than the bit rate of the optical signalsbeing switched. So, with this approach there is a bigFigure 2: Types of Multistage NetworksB. Comparison of Electronic and Optical NetworksThere are lots of benefits of optical networks overthe electronic ones. The main benefit of the opticalnetworks over the electronic network is the high speed ofthe Optical signals. In the optical networks light istransmitted which has a very good speed but in theelectronic Multistage interconnection networks electricityis used which has very slow speed. The second advantageis the Bandwidth. Now a day’s applications incommunication require high bandwidth, the opticalnetworks gives combination of very high bandwidth andlow latency. Therefore, they have been used in theparallel processing applications. Optical MINs are alsoused in wide area networks, which require less errorprobability and very high bandwidth. Fiber optic© 2010 ACADEMY PUBLISHER

144 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010transmission distance is significantly greater than theelectronic ones, signal need not to be regenerated inoptical networks. Optical fiber has very less weight incomparison to electronic MINs. Thus Optical networksgive the combination of high bandwidth and low latency.TABLE IICOMARISON OF ELECTRONIC AND OPTICAL NETWORKSCharacteristicsElectronicMultistageNetwotksOpticalMultistageNetworksSpeed Less HighEnergyElectricity LightTransmittedBandwidth Used for lessbandwidthapplicationsUsed for highbandwidthapplicationsLatency High LessError Probability High LessWeight More LessCost Less MoreSwitching Packet Switching CircuitSwitchingPathProvide Multi path Provide singlefrom source to path fromdestination. source todestinationComplexity More Complex Less ComplexStructureconsidered2-dimensional 3-dimensionalIV. PROBLEMS IN OPTICAL NETWORKSDue to the difference in speeds of the electronic andoptical switching elements and the nature of opticalsignals, optical MINs also hold their own challenges.A. Path Dependent LossPath dependent loss means that optical signalsbecome weak after passing through an optical path. In alarge MIN, a big part of the path-dependent loss isdirectly proportional to the number of couplers that theoptical path passes through [16]. Hence, it depends on thearchitecture used and its network size. Hence, if theoptical signal has to pass through more no of stages orswitches the path dependent loss will be more.B. Optical CrosstalkOptical crosstalk occurs when two signal channelsinteract with each other. There are two ways in whichoptical paths can interact in a switching network. Thechannels carrying the signals could cross each other.Alternatively; two paths sharing a switch couldexperience some undesired coupling from one path toanother within a switch. For example, assume that thetwo inputs are y and z, respectively, the two outputs willhave ly+lxz and lz+lxy, respectively, where l is path lossand x is signal crosstalk in a switch. Using the best devicex=35 dB and l=0.25 dB. For more practically availabledevices, it is more likely that x=20 dB and l=1 dB [5].Hence, when a signal passes many switches, the inputsignal will be distorted at the output due to the loss andcrosstalk introduced on the path.Crosstalk problem is more dangerous than the pathdependentloss problem with current optical technology.Thus, switch crosstalk is the most significant factor thatreduces the signal-to-noise ratio and limits the size of anetwork. Luckily, ensuring that a switch is not used bytwo input signals simultaneously can eliminate first-ordercrosstalk. Once the major source of crosstalk disappears,crosstalk in an optical MIN will have a very small effecton the signal-to-noise ratio and thus a large optical MINcan be built and effectively used in parallel computingsystems.V. APPROACHES TO SOLVE CROSSTALKA. Space Domain ApproachOne way to solve crosstalk problem is a spacedomain approach, where a MIN is duplicated andcombined to avoid crosstalk [8]. The number of switchesrequired for the same connectivity in networks with spacedomain approach is slightly larger than twice that for theregular network. This approach uses more than doublethe original network hardware to achieve the same. Thusfor the same permutation the hardware or we can say theno of switches will be double. Thus cost will be morewith the networks using space domain approach. In allthe four cases only one input and only one output isactive at a given time so that no cross talk occurs. Withthe space domain approach, extra switching elements(SEs) and links are used to ensure that at most one inputand one output of every SE will be used at any giventime.Figure 3. Crosstalk avoidance using space domain approachB. Time Domain ApproachAnother way to solve the problem of crosstalk is thetime domain approach [3]. With the time domainapproach, the same objective is achieved by treatingcrosstalk as a conflict; that is, two connections will beestablished at different times if they use the same SE.Whereas we want to distribute the messages to be sent tothe network into several groups, a method is used to find© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 145out which messages should not be in the same groupbecause they will cause crosstalk in the network. A set ofconnections is partitioned into several subsets such thatthe connections in each subset can be establishedsimultaneously in a network. There is no crosstalk inthese subsections. This approach makes importance inoptical MINs for various reasons. First, most of themultiprocessors use electronic processors and opticalMINs. There is a big mismatch between the slowprocessing speed in processors and the highcommunication speed in networks carrying opticalsignals [15]. Second, there is a mismatch between therouting control and the fast signal transmission speed. Toavoid crosstalk, the TDM approach is used, where the setof messages are partitioned into several groups such thatthe messages in each group can be sent simultaneouslythrough the network without any crosstalk.VI. METHODS FOR MESSAGE PARTITIONING INTDM APPROACHA. Window methodWindow method is the method that is used to find themessages that are not in the same group because it causescrosstalk in the network. If we consider the network ofsize N x N, there are N source and N destination address.Combining source and its destination address formscombination matrix. From this, optical window size is M-1 where M = log 2 N and N is size of network. In windowmethod, the number of windows is equal to the number ofstages [11].After finding conflicts using window method,conflict graph is generated shown in figure. The numberof nodes is the size of the network. The nodes that arehaving conflict are connected through edge. Degree ofeach message is the number of conflicts to the othermessage. Conflict graph is shown in figure 4.Figure 4. Conflict graphThe conflict matrix is a square matrix with N x Nentry, it consists of the output of the window method, asshown in figure 5. The definition of Conflict Matrix is thematrix M ij with size N x N. N is the size of the network.B. Improved window methodIn this method the first window is eliminated for thiswe make the conflict matrix initialized to 0, here Numberof windows is M-1. It takes less time to find conflictsthan the windows method. Therefore, it is calledimproved windows method [11,12].Figure 5. Conflict matrix© 2010 ACADEMY PUBLISHER

146 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Figure 7. Bitwise window methodC. Bitwise combination matrixFigure 6. Improved window methodFor Bitwise combination matrix, all binary bits ofsingle rows is each windows are converted to decimal,no. of window is reduced to n. By this method, time isreduced approximately by ten times. It is very effectivemethod even when the network is very large.For example the bitwise combination matrix for 8 8network size is demonstrated in Fig. 5. The number ofcolumns in WM is 6 (Ci, i = 2n) and for bitwise WM is 2(Ci, i = n) [11,12].VII. CONCLUSIONIn this paper properties of electronic and opticalMINs have been explained and compared. It isconcluded that for today’s applications such as in widearea networks (WANs) optical networks is the promisingchoice to meet the high demand of Speed and Bandwidth.The paper also describes the problems and solutions ofoptical MINs. Various methods available in literature,which are used for crosstalk avoidance in TDM approach,are described in detail. It has been observed that theimproved window method takes lesser time to findconflicts as compared to window method. The bitwisewindow method reduces the execution time ten timesthan the other algorithms even when the network size islarge.REFERENCESCombination MatrixD. Bitwise window methodBitwise Combination MatrixIn this method, source and destination address is indecimal format. Number of windows is log 2 N. Thus, fromcombination matrix, the optical window size is only onefor a different network size and the number of window islog 2 N. In other words, there are only one decimal numberin each row and each window for comparison and findinga conflict [11,12].[1] A. Verma and C.S. Raghvendra, “InterconnectionNetworks for Multiprocessors and Mul-ticomputers:Theory and Practice”, IEEE Computer Society Press,Los Alamitos, California, 1994.[2] A.K. Katangur, Y. Pan and M.D Fraser, “MessageRouting and Scheduling in Optical Mul-tistageNetworks Using Simulated Annealing”, InternationalProceedings of the Parallel and Distributed ProcessingSymposium (IPDPS), 2002.[3] C. Qiao, and R. Melhem, “A Time Domain ApproachFor Avoiding Crosstalk In Optical Blocking Multi-stageInterconnection Networks”, Journal of LightwaveTechnology, vol. 12 no. 10, October 1994, pp. 1854-1862.[4] C. Siu and X. Tiehong , “New Algorithm for MessageRouting and Scheduling in Optical MultistageInterconnection Network”, Proceedings of InternationalConferwnce on Optical Communications Systems andNetworks, 2004.[5] D. K. Hunter and I. Andonovic, “Guided wave opticalswitch architectures,” International Journal ofOptoelectronics, vol. 9, no. 6, 1994, pp. 477-487.[6] Hasan, “Rearrangeability of (2n − 1)-Stage Shuffle-Exchange Networks”, Society for Industrial and AppliedMathematics, vol. 32, no. 3, 2003, pp. 557-585.[7] J. T. Blaket and K.S. Trivedi, “Reliabilities of TwoFault-Tolerant Interconnection Networks”, Proceedingof IEEE, 1988, pp. 300-305.[8] K. Padmanabhan and A. Netravali, “Dilated Networksfor Photonic Switching”, IEEE Transactions onCommunication, vol. 35, no. 12, 1987, pp. 1357-1365.[9] L. N. Bhuyan and D.P. Aggarwal, “Design andperformance of generalized interconnection networks”,© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 147IEEE Transactions on computers, vol. C-32, no. 2,1983, pp. 1081–1090.[10] L. N. Bhuyan, Q. Yang Qing and D.P. Aggarwal,“Performance of Multiprocessor InterconnectionNetworks”, IEEE Computers, vol. 22, 1989, pp. 25-37.[11] M. A, M. Othman and R. Johari, “An efficient approachto avoid crosstalk in optical Omega Network”,International Journal of Computer, Internet andManagement, vol. 14, no. 1, 2005, pp. 50-60.[12] M. Ali, M. Othman, R. Johari and S. Subramaniam,“New Algorithm to Avoid Cros- stalk in OpticalMultistage Interconnection Networks”, Proceedings ofIEEE International Conference on Network (MICC-ICON), 2005, pp. 501-504.[13] M. Fang, Layout Optimization for Point to Multi Point,Wireless Optical Networks via Simulated Annealing &Genetic Algorithm”, Master Project, University ofBridgeport, 2000.[14] Regis Bates, “Optical Switching and NetworkingHandbook”, McGraw-Hill, New York, 2001.[15] X. Shen, F. Yang and Y. Pan, “Equivalent permutationcapabilities between time division optical omeganetworks and non-optical extra-stage omega networks”,IEEE Transactions on Networking, vol.9, no. 4, 2001,pp. 518-524.[16] Y. Pan , X. Lin, and X. Jia, Evolutionary Approach ForMessage Scheduling In Optical Omega Networks, FifthIntern. Conf. on Algorithms and Architectures forParallel Processing (ICA3PP), 2002.Himanshu Aggarwal, Ph.D., isReader in Computer Engineering atUniversity College of Engineering,Punjabi University, Patiala. He hasmore than 16 years of teachingexperience and served academicinstitutions such as Thapar Instituteof Engineering & Technology,Patiala, Guru Nanak DevEngineering College, Ludhiana and Technical Teacher'sTraining Institute, Chandigarh. He is an active researcher whohas supervised many M.Tech. Dissertations and contributed 32articles in Conferences and 13 papers in research Journals. Hisareas of interest are Information Systems, ERP and ParallelComputing.BiographiesRinkle Rani Aggarwal,B.Tech (Computer Science &Engg.), M.S. (SoftwareSystems), is Senior Lecturer inComputer Science &Engineering Department atThapar University, Patiala. Shehas more than 12 years ofteaching experience and servedacademic institutions such as Guru Nanak DevEngineering College, Ludhiana and S.S.I.E.T 'Derabassi.She has supervised many M.Tech. Dissertations andcontributed 23 articles in Conferences and 11 papers inresearch Journals. Her areas of interest are ParallelComputing and Algorithms.Lakhwinder Kaur, Ph.D. isReader in Computer Engineering atUniversity College of EngineeringPunjabi University, Patiala. Shehas 17 years of teachingexperience. She has published 12research papers in InternationalJournals. Her areas of interest areImage processing, ParallelComputing and ComputerGraphics.© 2010 ACADEMY PUBLISHER

148 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Web Based Hindi to PunjabiMachine Translation SystemVishal Goyal and Gurpreet Singh LehalDepartment of Computer Science, Punjabi University, Patiala, Punjab, India{vishal.pup,gslehal}@gmail.comAbstract - Hindi and Punjabi are closely related languageswith lots of similarities in syntax and vocabulary BothPunjabi and Hindi languages have originated from Sanskritwhich is one of the oldest language. In terms of speakers,Hindi is third most widely spoken language and Punjabi istwelfth most widely spoken language. Punjabi language ismostly used in the Northern India and in some areas ofPakistan as well as in UK, Canada and USA. Hindi is thenational language of India and is spoken and used by thepeople all over the country. In the present research, BasicHindi to Punjabi machine translation system using directtranslation approach has been developed. The results ofthis translation system are surprisingly good. The systemincludes lexicon based translation, transliteration andcontinuously improving the system through machinelearning module. It also takes care of basic word sensedisambiguation.Index Terms - Machine Translation System, Closely relatedLanguages, Hindi, Punjabi, Natural Language Processing,Computational Linguistics, Transliteration.I. INTRODUCTIONMT is a field of research that has been around sincethe birth of electronic computers. Warren Weaver, adirector of the Rockefeller Foundation received muchcredit for bringing the concept of MT to the publicwhen he published an influential paper on usingcomputers for translation in 1949. MT is the name forcomputerized methods that automate all or part of theprocess of translating from one human language toanother. Fully-automatic general purpose high qualitymachine translation system (FGH-MT) is extremelydifficult to build. In fact, there is no system in theworld of any pair of languages which qualifies to becalled FGH-MT. This paper explains the methodologyfollowed for developing the machine translationsystem between closely related languages – Hindi andPunjabi. Closely related languages have lots ofsimilarities in syntax and vocabulary. MachineTranslation between closely related languages is easierthan between language pairs that are not related witheach other. Having many parts of their grammars andvocabularies in common reduces the amount of effortneeded to develop a translation system between relatedlanguages. Closely related languages are the languages ofpeople who have similar cultures and common historicalroots.II. HISTORYThe first attempt to verify the hypothesis that relatedlanguages are easier to translate started in mid 80s atCharles University in Prague [FEMTI; Hajic et al. 2000].The project was called RUSLAN and aimed at thetranslation of documentation in the domain of operatingsystems for mainframe computers. From that date to tilldate so many examples are there in history which supportthe argument that with close languages, the quality ofMT system, with simple techniques, is better. To name afew one are CESILKO (a system for translating Czechand Slovak), MT system for translating Turkish ToCrimean Tatar etc. We are also trying to strengthen thesame concept by experimenting with a word for worddirect translation system for Hindi to Punjabi. Theselanguages are very closely related and have manyfeatures in common.III. SYSTEM DESCRIPTIONThe major task behind direct Machine translationsystem is developing an exhaustive lexicon consisting ofsource language words along with its correspondingtranslated version of the target language. There is nomachine readable dictionary available for Hindi toPunjabi language. Two traditional dictionaries – onefrom Bhasha Vibhag, Patiala and another from NationalBook Trust has been published. We got it digitized andmoulded required for machine translation purpose whichis itself a big job. Now lexicon consists of approx.1,00,000 words. This lexicon is used for word for wordtranslation. Extending the lexicon demands large hindicorpus. Sometimes it is available in Unicode format andsometimes it is available in number of other non standardfonts like Susha, Agra, Krutidev etc. which needs to beconverted into Unicode first. Conversion is being donethrough Font Converter that converts non standard fontsinto Unicode Format. The beauty of the developed Fontconverter is that it is able to convert MS-Access files, MSWord files, HTML files and Text Files. Because thecorpus can be in any file type, so it handles all thepossible file types. Besides translation it performs thetask of transliteration as well. Transliteration means toreplace character by character of word from sourcelanguage character to target language character likeेमचंद into ਪੇਮਚਂਦ. This system is basically an extension© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.148-151

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 149of previous system that does not handle any word sensedisambiguation. The above said system just checks thewords to be translated in the dictionary, if found it isreplaced with the translated version stored in thedictionary, otherwise it is transliterated. But now in theextended system, the lexicon is divided into two parts –one table consists of words with no disambiguation andsecond consists of words that have multiple meaningsdepending upon the context of the word in which it hasbeen used in the sentence.IV. SYSTEM ARCHITECTUREThe architecture for the HPMTS (Hindi to PunjabiMachine Translation System) consists of number ofmodules that are listed below:a. Training the system with training corpusb. Input Text Font Conversion into UnicodeFormatc. Hindi Text Normalizationd. Finding and Replacing Collocationse. Finding and replacing named entitiesf. Word to word translation using lexiconsg. Resolving Ambiguity among wordsh. Transliteration of wordsi. Post Processingj. Improving the accuracy of the system throughmachine learning during every translation job.k. Testing the system using test corpus other thantrain corpusIn the above architecture, the most important part andstarting point is to train the system. Train the systemmeans generating the lexicon using the already existingcorpus. The second module is optional and is skipped ifthe inputted text is already in Unicode format. UnicodeFont requirement arises due to internalization of thesystem and making the system free from specific fontdependency. This font converter can be also used forconverting the non-Unicode corpus into Unicode formatcorpus. Indian language words face spellingstandardization issues, thereby resulting in multiplespelling variants for the same word. The main reason forthis phenomenon can be attributed to the phonetic natureof Indian Languages and multiple dialects. To give anidea of this data problem, these words were found –मंिजल, मिजल, मंिज़ल Third module is Hindi TextNormalization that solves this spelling variant problem.Hindi text is normalized into standard spellings before itgoes for translation. Next Module of the system find andreplaces all the collocations using the lexicon enteries. ACollocation is an expression consisting of two or morewords that correspond to some conventional way ofsaying things. Or in the other words of Firth (1957:181) :“Collocations of a given word are statements of thehabitual or customary places of that word”. This modulehelps in increasing the accuracy of the translation.Generating Lexicon for Collocations is itself achallenging task. Then comes the turn of the heart of thesystem – word for word translation uses the lexicon. Thissearch for the Hindi word in the lexicon and replaces itwith the corresponding Punjabi translated version presentin the lexicon. If this Hindi word is not found in thelexicon it searches that word in the database ofambiguous words, if found using tri-gram approach itresolves the ambiguity of word and replaces it withcorrect Punjabi meaning among multiple Punjabimeanings. For Example, the hindi word सरूप can betranslated into either of the two Punjabi words - ਸਮਾਨ,ਸੁ ੰ ਦਰ. But how will the system decide which word tochoose is basically to know the context in which theHindi word सरूप has been used in the sentence. If theword is not found in both the tables it means it is notavailable in the database and need to be transliterated.For improving the accuracy of the system, this is must toknow the system about which new words have beencome across and if they have been transliteratedaccurately or not. If they were not present in the databaseand need to be present, it is added to lexicon for futuretranslations. If it has been translated wrongly butrequired one, it is corrected first before adding to thelexicon. In this way this is the ongoing improvement ofthe system performance during every translation exercisethrough machine learning module. Post ProcessingModule takes into consideration some commongrammatical mistakes that has been done duringtranslation phases and based on the rules framed, itremoves those mistakes and increases the accuracy to thesystem. Now system has been trained a lot by number oftranslation exercises, it is time to check the accuracy ofthe system by testing the system through test data otherthan the data used for training. Testing the system is alsovery tedious task. First step in it is to prepare the testcases that covers all the possibilities.V. WEB BASED TOOLResearch must not be restricted to papers, It must bepropagated to public for use and test. Taking this aim, thewhole system has been developed as a web tool and isonline for use fre of cost. The website address ishttp://h2p.learnpunjabi.org/ . Following are the featuresof this web tool:a. Hindi Text can be written in Unicode encodingby using the most popular Hindi Font Krutidev.This concept is very useful for those who are inhabit of typing the text in Krutidev and laterthey find some source for converting them intoUnicode encoding. Thus, this feature has solvedtheir purpose in very easy manner. Now, theywill type in their style and the typed matter willalso be in Unicode.b. The text can also be input to the system fortranslation through text file. File can be readusing the Browse button provided.c. Input text can be translated into Punjabi text byjust clicking the Translate button. Withinseconds, the text is translated into Punjabi.© 2010 ACADEMY PUBLISHER

150 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010d. Input Text can be transliterated also, if there isneed by clicking on the Transliterate button.e. Email facility has also been provided. Text canbe typed in Hindi and English both. Subject canbe written in both the languages. Then whilesending email there is an option of sending theemail in original or after translating in Punjabi.This is also very powerful feature.f. The main feature of the webtool is user cantranslate the full Hindi website into Punjabiwebsite by just providing the link of hindiwebsite and press Translate. The hindi websiteis translated into Punjabi within seconds andPunjabi version of the website is displayed inthe same format as it was originally in hindiwebsite.VI. SCREEN SHOTSFollowing is the screen shot of web based machinetranslation system:A. Input Text:VII. SAMPLE OUTPUTभारत और पड़ोसगुवार, 01 नवंबर, 2007 को 08:02 GMT तक के समाचारकनाटक को लेकर राजनीितक सरगम बढ़कु मारःवामी ने राजनीितक चाल बदलते हुए येदयुरपा कोमुयमंऽी के प म समथन देने का फ़ै सला कया हैकनाटक को लेकर राजनीितक गहमागहमी तेज़ हो गई है औरअब यह मामला दली आ गया है.संभावना है क जद ह कनाटक के राजनीितक भवंय काकोई फ़ै सला हो जाएगा.बुधवार को जहाँ भाजपा के उच ःतरय ूितिनिध मंडल नेूधानमंऽी मनमोहन िसंह से मुलाक़ात क है ूधानमंऽी औरसोिनया गाँधी ने कांमेस कोरमुप क बैठक म कनाटक केराजनीितक हालात पर चचा क है.उधर भाजपा ने धमक द है क यद भाजपा-जनतादल (एस)को सरकार बनाने के िलए आमंऽत नहं कया गया तो दोनदल के 129 वधायक रापित के सामने परेड करगे.उलेखनीय है क जनता दल (एस) और भाजपा का गठबंधनसात अू बर तक सा म था. लेकन समझौते के अनुसारमुयमंऽी बदले जाने के ववाद के चलते गठबंधन टूट गयाऔर भाजपा ने सरकार से समथन वापस ले िलया.B. Translated Output Text:ਭਾਰਤ ਅਤੇ ਗੁਆੰ ਢਵੀਰਵਾਰ , 01 ਨਵੰ ਬਰ , 2007 ਨੂ ੰ 08 : 02 GMT ਤੱਕ ਦੀਆਂਖ਼ਬਰਕਰਨਾਟਕ ਨੂ ੰ ਲੈਕੇ ਰਾਜਨੀਤਕ ਜੋਸ਼ ਵਿਧਆਕੁਮਾਰਸਵਾਮੀ ਨ ਰਾਜਨੀਤਕ ਚਾਲ ਬਦਲਦੇ ਹੋਏ ਯੇਿਦਉਰੱਪਾ ਨੂ ੰਮੁੱਖਮੰ ਤਰੀ ਦੇ ਰੁਪ ਿਵੱਚ ਸਮਰਥਨ ਦੇਣ ਦਾ ਫੈਸਲਾ ਕੀਤਾ ਹੈਕਰਨਾਟਕ ਨੂ ੰ ਲੈਕੇ ਰਾਜਨੀਤਕ ਗਿਹਮਾਗਿਹਮੀ ਤੇਜ਼ ਹੋ ਗਈ ਹੈ ਅਤੇਹੁਣ ਇਹ ਮਾਮਲਾ ਿਦੱਲੀ ਆ ਿਗਆ ਹੈ .ਸੰ ਭਾਵਨਾ ਹੈ ਿਕ ਜੱਲਦੀ ਹੀ ਕਰਨਾਟਕ ਦੇ ਰਾਜਨੀਤਕ ਭਿਵੱਖ ਦਾਕੋਈ ਫੈਸਲਾ ਹੋ ਜਾਵੇਗਾ .ਬੁੱਧਵਾਰ ਨੂ ੰ ਿਜੱਥੇ ਭਾਜਪਾ ਦੇ ਚ ਪੱਧਰ ਪਿਤਿਨੱਧੀ ਮੰ ਡਲ ਨਪਧਾਨਮੰ ਤਰੀ ਮਨਮੋਿਹਨ ਿਸੰ ਘਨਾਲ ਮੁਲਾਾਤ ਕੀਤੀ ਹੈਪਧਾਨਮੰ ਤਰੀ ਅਤੇ ਸੋਿਨਆ ਗਧੀ ਨ ਕਗਰਸ ਕੋਰਗਰੁਪ ਦੀ ਬੈਠਕਿਵੱਚ ਕਰਨਾਟਕ ਦੇ ਰਾਜਨੀਤਕ ਹਾਲਾਤ ਤੇ ਚਰਚਾ ਕੀਤੀ ਹੈ .ਓਧਰ ਭਾਜਪਾ ਨ ਧਮਕੀ ਿਦੱਤੀ ਹੈ ਿਕ ਜੇਕਰ ਭਾਜਪਾ - ਜਨਤਾਦਲ (ਏਸ ) ਨੂ ੰ ਸਰਕਾਰ ਬਣਾਉਣ ਲਈ ਸੱਿਦਆ ਨਹ ਕੀਤਾ ਿਗਆ ਤ ਦੋਨਦਲ ਦੇ 129 ਿਵਧਾਇਕ ਰਾਸ਼ਟਰਪਤੀ ਦੇ ਸਾਹਮਣੇ ਪਰੇਡ ਕਰਣਗੇ .ਿਲਖਣ ਯੋਗ ਹੈ ਿਕ ਜਨਤਾ ਦਲ ( ਏਸ ) ਅਤੇ ਭਾਜਪਾ ਦਾ ਗੰ ਢ-ਜੋੜਾਸੱਤ ਅਕਤੂਬਰ ਤੱਕ ਸੱਤਾ ਿਵੱਚ ਸੀ . ਪਰ ਸਮੱਝੌਤੇ ਦੇ ਅਨੁਸਾਰਮੁੱਖਮੰ ਤਰੀ ਬਦਲੇ ਜਾਣ ਦੇ ਝੱਗੜਾ ਦੇ ਚੱਲ ਦੇ ਗੰ ਢ-ਜੋੜਾ ਟੁੱਟ ਿਗਆਅਤੇ ਭਾਜਪਾ ਨ ਸਰਕਾਰ ਨਾਲ ਸਮਰਥਨ ਵਾਪਸ ਲੈ ਿਲਆ .The accuracy of the translation comes out be approx.95%.VIII. CONCLUSIONThe present system is translating any complexsentence. The System accuracy is measured up to 95%.This web tool has number of applications in real world.This web tool can be used by Newspaper agencies, anywebsite owner, using email facilty by community ofdifferent countries or regions i.e. Writing the email inHindi and recipient will receive the email in Punjabi.Thus removing the language bars, communicationbecomes easy in one’s owm language.© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 151REFERENCES1. Kemal Altintas, Dept. of Computer Engineering, BilkentUniversity, Ankara, Turkey, “A Machine TranslationSystem Between a Pair of Closely Related Languages”.Internet:http://www.cs.bilkent.edu.tr/~ilyas/PDF/iscis2002.pdf2. Bharati A., Chaitanya V and Sangal R, "Natural Languageprocessing: A Paninian Perspective", Prentice Hall of India,New Delhi, 1995.3. Joseph Seasly, “Machine Translation: A Survey ofApproaches”, University of Michigan, Ann Arbor, 2003.4. Durgesh Rao, “Machine Translation in India: A briefsurvey”, National Centre for Software Technology,Mumbai, 2001. Internet:http://www.eldra.fr/en/rproj/scalla/SCALLA2001Rao.pdf5. R.M.K. Sinha, R. Jain and A. Jain "Translation from Englishto Indian Languages: ANGLABHARTI Approach",Proceeding of STRANS-2002, pp. 69-85, 2002.6. Christopher D. Manning and Hinrich Schutze, “Foundationsof Statistical Natural Language Processing”, MITPress,1999.7. Manish Sinha, Mahesh Kumar Reddy, PushpakBhattacharya, “Hindi Word Sense Disambiguation”.Internet:http://www.cse.iiitb.ac.in/Pb/papers/HindiWSD.pdf8. R. Canals-Marote, A. Esteve-Guillén, A. Garrido-Alenda,M.I. Guardiola-Savall, A. Iturraspe-Bellver, S. Montserrat-Buendia, S. Ortiz-Rojas, H. Pastor-Pina, P.M. Pérez-Antón,and M.L. Forcada, “The Spanish_ Catalan machinetranslation system interNOSTRUM”, Internet:http://internostrum.com/docum/iN-MTS.pdf9. Kemal Altintas, Ilyas Cicekli,”A Machine TranslationSystem Between a Pair of Closely Related Languages”,Internet:http://www.cs.bilkent.edu.tr/~ilyas/PDF/iscis2002.pdf10. Kevin P. Scannell ,”Machine translation for closely relatedlanguage pairs” Internet: http://borel.slu.edu/pub/ga2gd.pdf11. Jan AJIC, Jan HRIC, Vladislav KUBON, “CESILKO – anMT system for closely related languages”, Internet:http://www.cs.ust.hk/acl2000/Demo/03_kubon.pdf© 2010 ACADEMY PUBLISHER

152 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010Protecting Data From the Cyber Theft – AVirulent Disease1 Dr. S.N. Panda and 2 Vikram Mangla1 Professor & Principal, 2 Assistant Professor1 RIMT-IMCT, Mandi Gobind Garh, Punjab.2 Chitkara Institue of Engineering & Technology, Rajpura, Punjab.1 panda.india@gmail.com, 2 mangla.vikram@gmail.comAbstract - Network security policies are essential elements inInternet security. Network security perimeter devices suchas firewalls, IPSec, and IDS/IPS devices operate based onlocally configured policies. Malware-related data breacheshave reached pandemic proportions as criminals discoverthat Internet crime is easy to commit, highly lucrative, andlargely under-policed. With a few hundred dollars, a cybercriminal can begin a career of breaking into computers tosteal identity and confidential data for sale to the highestbidder. This paper will cover current and emerging trends ofstealth malware, such as moving primarily to the Web sincemost organizations allow Web traffic into the network. Itwill also cover new advances in network securitytechnologies that use multi-phase heuristic and virtualmachine analysis to detect and mitigate the damages thatresult from malware-related data thefts.Index Terms - Network Security, Web Threats, Malware,PhishingI. INTRODUCTIONWith the global connectivity provided by the Internet,network security has gained significant attention inresearch and Industrial communities. Due to theincreasing threats of network attacks, network securitydevices such like firewalls and IPSec gatewaye havebecome important integrated elements not only inenterprise networks but also in small size and homenetworks. Motivated by the lure of profits from the sale ofstolen confidential information, cyber criminals today areshifting to the Web as their chosen attack vector, whichprovides an ideal environment for cyber crime. Malwarerelateddata breaches have reached pandemic proportionsas criminals discover that Internet crime is easy tocommit, highly lucrative, and largely under-policed. Witha few hundred dollars, a cyber criminal can begin a careerof breaking into computers to steal identity andconfidential data for sale to the highest bidder. Fraudsterswho purchase the data have developed a variety ofschemes to monetize that information ranging fromtransacting unauthorized stock trades to transferring fundsto offshore bank accounts. The cyber crime economy is sorobust that there is a vibrant market for professionalmalware toolkits available for $500 to $1,000 and comepre-configured with a range of attack modules, exploit‘maintenance’ updates, and 24 x 7 online technicalsupport.Many Web threats can be deployed unbeknownst to theuser, requiring no additional action than merely opening aWeb page. Large numbers of users, an assortment oftechnologies, and a complex network structure providecriminals with the targets, exploitable weaknesses, andanonymity required for large-scale fraud. Web threatspose a broad range of risks, including financial damages,identity theft, and loss of confidential businessinformation, theft of network resources, damaged brand orpersonal reputation, and erosion of consumer confidencein e-commerce. These high stakes, the pervasive use ofthe Web, and the complexity of protecting against Webthreats combine to form perhaps the greatest challenge toprotecting personal and business information in a decade.In August 2007, a scene played out as cyber criminalsinfiltrated the monster.com job site through “Monster forEmployers” accounts, compromising the personalinformation of 1.6 million users. Many of these users thenreceived official-looking emails, claiming to be frommonster.com and encouraging them to download a “helperapplication” that turned out to be yet more malware.These attacks were well-researched, using familiarlanguage and branding, and coded to transfer data slowly,under the radar of IT administrators looking for suspiciousnetwork traffic.[1] Web threats also include malware thatis downloaded from an email attachment, but accesses theWeb to convey information to the hacker. In 2007,fraudulent emails were sent purporting to be from theFederal Trade Commission. These emails claimed that acomplaint had been filed against the company andcontained an attachment. If the recipient opened theattachment, a keylogging Trojan was deployed thatattempted to steal login information from the user’scomputer and send it back to the hacker. [2].Phishing is a prevalent Web threat, spoofing legitimatecompanies to trick people into providing confidentialinformation. Consumer phishing is wide-spread, sendingemails that spoof organizations like banks and on-lineretailers. These phishing emails often use links to takerecipients to Web sites where confidential information isgathered. Employees can fall victim to these consumerthreats, but phishing can also affect corporations moredirectly. In 2005, phishing emails targeted CEOs andother high-level executives of US credit unions in anattempt to gain control of millions of personal financialrecords. The email messages contained a link to a Website where a Trojan was downloaded. Even one successful© 2010 ACADEMY PUBLISHERdoi:10.4304/jetwi.2.2.152-155

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 153infection could have caused millions of dollars of damageand caused irreparable harm to hundreds of thousands ofusers through identity and asset theft. [3]But Web threats don’t just steal confidentialinformation; they can also steal network resources.Variations of e-greeting card spam were sent throughout2007. These simple spam messages told recipients that afriend had sent them an e-greeting card and to follow thelink in the email to view the card. If recipients followedthe link, it took them to a Web site that downloadedmalicious code.This code hijacked the computer, turning it into a “bot”and allowing the hackers to use the machine for their ownpurposes—sending spam, hosting malicious Web sites,and much more. Consumer and corporate computers wereinfected by the millions. Hackers network these infectedcomputers to create botnets, stealing resources and furtherperpetuating their fraudulent activities.II. WEB THREATS DEFINEDWeb threats are any threat that uses the Web tofacilitate cyber crime. They are sophisticated in theirmethods, using multiple types of malware and fraud, all ofwhich utilize HTTP or HTTPS protocols, but can alsoemploy other protocols as components of the attack, suchas links in email or IM, or malware in attachments or onservers that access the Web. The creators of such threatsfrequently update Web site content, variants, and malwaretypes in order to evade detection and achieve greatersuccess.Web threats based on malware are hidden within Webpages and victims are infected when they visit the page.Fraudulent sites mimic legitimate business Web sites anduse social engineering to request visitors to discloseconfidential information. Individuals once characterizedas hackers, virus writers, spammers, and spy ware makersare now simply known as cyber criminals with financialprofit their primary aim.Over the last 15 years, information security threatshave evolved through a series of incarnations. In eachcase, malware writers and fraudsters sought out themedium that was most used and least protected (forexample email). Today, a new wave of threats is emergingthat uses the Web as a delivery vehicle. These Web threatsare gaining traction at a time when the Web has become amajor commerce engine as well as social networkingvehicle, with usage continuing to grow.At the same time, the Web is relatively unprotected,compared to messaging for example, as a medium todeliver malware and conduct fraud. According to IDC,“Up to 30% of companies with 500 or more staff havebeen infected as a result of Internet surfing, while only20%-25% of the same companies experienced viruses andworms from emails.” [4]III. WEB THREAT DELIVERY MECHANISMSWeb threats can be divided into two primarycategories, based on delivery method – push and pull.Push based threats use spam, phishing, or other fraudulentmeans to lure a user to a malicious (often spoofed) Website, which then collects information and/or injectsmalware. Push attacks use phishing, DNS poisoning (orpharming), and other means to appear to originate from atrusted source. Their creators have researched their targetwell enough to spoof corporate logos, official Web sitecopy, and other convincing evidence to increase theappearance of authenticity. Precisely-targeted push-basedthreats are often called “spear phishing” to reflect thefocus of their data gathering (“phishing”) attack.Spear phishing typically targets specific individualsand groups for financial gain. In November 2006, amedical center fell victim to a spear phishing attack.Employees of the medical center received an email tellingthem they had been laid off. The email also contained alink that claimed to take the recipient to a careercounseling site. Recipients that followed the link wereinfected by a keylogging Trojan. [5] In other push-basedthreats, malware authors use social engineering such asenticing email subject lines that reference holidays,popular personalities, sports, pornography, world events,and other popular topics to persuade recipients to open theemail and follow links to malicious sites or openattachments with malware that accesses the Web.Pull-based threats are often referred to as “drive-by”threats, since they can affect any visitor, regardless ofprecautions. Pull threat developers infect legitimate Websites, which unknowingly transmit malware to visitors oralter search results to take users to malicious sites. Uponloading the page, the user’s browser passively runs amalware downloader in a hidden HTML frame (IFRAME)without any user interaction. Both push- and pull-basedWeb threat variants target infection at a regional or locallevel (for example, via local language sites aimed atparticular demographics), rather than using the massinfection technique of many earlier malware approaches.These threats typically take advantage of Internet port 80,which is almost always open to permit access to theinformation, communication, and productivity that theWeb affords to employees.IV. TODAY’S INSIDER - THREAT IS STEALTH MALWARELaw enforcement, computer crime experts, and eventhe military are playing catch up to the threat posed toconsumers, businesses, and national security as cybercriminals cash in on stolen identity data, fraudulent onlinetransactions, and cyber espionage. It is no surprise that therise in cyber crime has coincided with the increased use ofthe Internet and especially “Web 2.0” technologies.Web sites and applications now support usercontributedcontent, syndicated content, iframes, thirdpartywidgets (or applets), and convoluted advertisingdistribution networks into which ‘stealth’ malware caneasily be injected somewhere along the line. In a 2007USENIX paper, Google researchers determined thatapproximately 9% of all suspicious web sites launched“drive-by” downloads of stealth malware binaries[12].Government studies[13] estimate that 65% of all exploits© 2010 ACADEMY PUBLISHER

154 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010now enter via the Web and IBM Internet Security Systems(ISS) estimates that nearly 100% of Web attacks nowutilize obfuscated JavaScript as a very effective techniqueto bypass antivirus and intrusion prevention.Today, once a PC is infected with stealth malware, ittypically opens two-way communications to a “commandand control” (C&C) server to establish a channel back tothe cyber criminal. This allows the “bot” (as in “robotcomputer”) to report status as well as any valuableinformation that is immediately accessible. Groups ofthese remotely controlled, malware-infected computersare commonly called botnets, and serve as the foundationof most cybercrime on the Internet.How do victims get infected? A user may be drawn bya phishing e-mail to a Web site hosted on a hijackedserver, which serves up a browser exploit; this downloadsand installs a bot on the user’s PC. The bot thendownloads more malware like “keyloggers” that silentlyrecord keyboard and mouse activities to execute furthercriminal activities, such as stealing user credentials andcapturing other sensitive information. All of this takesplace without the knowledge of the user or administrator.As their prevalence has increased, remote-controlmalware/botnets have become serious concerns forsecurity administrators.The recent January, 2009 malware-related data thefts atHeartland Payment Systems and earlier malwareRecent Research[14] Has Found:11 % of the world’s computers are enmeshed in at leastone botnet23 % of home computers become infected despite havingsecurity enabled72 % of corporate networks larger than 100 PC’s have aninfectioninfiltrations at Hannaford Supermarkets, University ofFlorida Medical Center, and NASA underscore theescalating threat of malware-related data breaches. TheIdentity Theft Resource Center, a nonprofit group focusedon understanding and preventing identity theft, reportedthat 656 known security breaches had taken place in 2008,reflecting a 47 percent increase over 2007’s total. As ofMarch 17, 2009 the resource center had already reported110 breaches in 2009.V. STEALTH MALWARE ATTACKS ARE OUTMANEUVERINGCONVENTIONAL DEFENSESDefending corporate networks from today’s malwarerelateddata thefts requires modern protection that goesbeyond current signature- and heuristic-based detectiontechniques. Modern threats exploit the inability ofconventional network protection to provide a unifieddefense against a criminal who attacks on multiple fronts,from OS and browser vulnerabilities to social engineering.The anachronistic concept of detecting infections with asingle technique, such as signatures, has left manybusinesses and consumers open to attack, despite theirdeployment of antivirus and IPS (intrusion preventionsystems). The sheer volume and escalating danger ofmodern attacks are overwhelming limited IT resourcesand outmaneuvering conventional defenses that mayalready be in place. To enable a more efficient IT securityprocess, accurate and timely identification of infectedmachines is the first step in preventing malware-relateddata breaches. And, the only viable solutions are thosethat provide thorough coverage across the many vectorsthat are used in attacks.VI. CONVENTIONAL APPROACHES FAIL TO PROTECTAGAINST WEB THREATSWeb threat scanning has specific requirements that arenot met by the traditional approach to virus scanning.Conventional antivirus software installed on clientmachines, for example, while crucial to the protection ofthese machines from a variety of threats, does notadequately protect against the evolving set of Web threats.One reason is that the conventional approach to virusprotection involves collecting samples of viruses,developing patterns, and quickly distributing thesepatterns to users. Because many Web threats are targetedattacks and span many variants, collecting samples isalmost impossible.The large numbers of variants use multiple deliveryvehicles (for example, spam, instant messaging, and Websites), rendering the conventional sample collection,pattern creation, and deployment process insufficient.Another reason that conventional virus detectionprocesses fall short involves a fundamental differencebetween these viruses and evolving Web threats.Conventional viruses were fundamentally designed tospread as quickly as possible, and were therefore ofteneasy to spot. With the advent of Web threats, malware hasevolved from this outbreak model to stealthy “sleeper”infections that are therefore difficult to detect viaconventional antivirus techniques.Recovering from infections also presents newchallenges. In some cases, Web threats may result in asystem infection that is so extensive (for example, via arootkit in which the system file is replaced) thatconventional uninstall or system cleaning approachesbecome useless. Infected systems often require a completesystem recovery, in which the hard drive is wiped and theoperating system, applications, and user data arereinstalled.VII. FUTURE WORKA New Approach Is Needed: Integrated, Multi-Layered Protection - Clearly, users need a newapproach to addressing Web threats that complementsexisting techniques.The most effective approach willemploy multiple layers of protection and incorporate arange of protective measures. In addition, the evolvingnature of the threat necessitates some form of informationfeedback and integration, in which information gatheredin one portion of the protection network is used to updateinformation in other layers. Any effective approach© 2010 ACADEMY PUBLISHER

JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 2, NO. 2, MAY 2010 155should also address all relevant protocols, because Webthreats leverage multiple protocols in their attacks, inparticular email as the initial delivery mechanism and theWeb as the threat host. However, other mechanisms canalso help perpetrate attacks such as links in IM andinfected files.Coordinating measures requires efficient, centralizedmanagement of region-specific expertise to help addressthe regional, and even localized nature of many of thethreats. The key to effectively addressing Web threats is amulti-layered approach. The network points arecategorized in four different layers (see Figure 2): 1) “inthe-cloud”(i.e. before the traffic reaches the Internetgateway), 2) at the Internet gateway, 3) across the networkservers, 4) and at the endpoint (for example, the client). Inthe below example, the description uses the points in thenetwork for high level organization and describes theprotocol protection and security technologies that can bedeployed at these points. The subsections on protocolprotection and security technologies describe emailsolutions first, which is often the first step in a Web threatattack, followed by Web solutions that directly protectWeb usage.A multi-layered approach is needed to protect againstthe broad range of Web threatsDNA of an Ideal Solution:Dynamic, real-time detection of threat: Finds thelatest stealth, 0-day attacksAccurate detection: No false positives, and no falsenegativesReturn on security investment: Easy to install,manage, support and scaleVIII. CONCLUSIONWeb threats are prevalent today and are growing innumbers and impact. Their complexity, large number ofvariants, and use of multiple vectors, combined with theirexploitation of the most commonly used medium today -the Web - make Web threats the most challenging threatthat consumers, businesses, and services providers, havefaced in a long time.Potential costs associated with these threats includeconfidential information leakage and theft of networkresources, with the adverse impact of erosion ofcustomers, trust, and brand reputation; regulatory andlegal implications; negative public relations; and loss ofcompetitive advantage. Because conventional approachesfail to protect against Web threats, the informationsecurity industry is at a crossroads. Businesses of all sizes,as well as service providers, need to deploy solutions viaan integrated, multi-layered approach to provide real-time,comprehensive protection against these threats.REFERENCES1. Gregg Keizer, Computerworld, August 19, 2007, “Identityattack spreads; 1.6M records stolen from Monster.com,”http://computerworld.com/action/article.do?command=viewArticleBasic&articleId=9031418&pageNumber=1.2. Dan Kaplan, SC Magazine, October 30, 2007, “FTC SpamContains Keylogging Trojan”,http://www.scmagazineus.com/FTC-spam-containskeylogging-trojan/article/58273/3. Paul F. Roberts, eWeek.com, December 16, 2005, “SpearPhishing Attack Targets Credit Unions,”http://www.eweek.com/article2/0,1895,1902896,00.asp.4. IDC, press release, July 18, 2006, “Private Internet Use byStaff Threatens IT Security in Danish Companies, SaysIDC,”http://www.idc.com/getdoc.jsp?containerId=pr2006_07_14_125434.5. Cara Garretson, NetworkWorld.com, January 11, 2006,“Spam that Delivers a Pink Slip”http://www.networkworld.com/news/2006/110106-spamspear-phishing.html6. Gregg Keizer, TechWeb Technology News, January 24, 2006,“Botnet Creator Pleads Guilty, Faces 25 Years,”http://www.techweb.com/wire/security/177103378.7. Marius Oiaga, Softpedia, October 4, 2006, “Hacking RussianTrio Gets 24 Years in Prison,”http://news.softpedia.com/news/Hacking-Russian-Trio-Gets-24-Years-in-Prison-37149.shtml.8. Byron Acohido and Jon Swartz, USA TODAY “Cybercrimeflourishes in online hacker forums,” October 11, 2006,http://www.usatoday.com/tech/news/computersecurity/infotheft/2006-10-11-cybercrime-hackerforums_x.htm.9. Police of the City of Munich, August 25, 2006,http://www.sueddeutsche.de/,tt3m3/muenchen/artikel/612/83529.10. Avivah Litan, “Phishing Attacks Escalate, Morph, and CauseConsiderable Damage,” Gartner, December 12, 2007.11. Tom Krazit, Cnet, “Two in three retail PCs are notebooks,”December 20, 2006,http://news.com.com/Two+in+three+retail+PCs+are+notebooks/2100-1044_3-6144921.html.12. Niels Provos, Dean McNamee, Panayiotis Mavrommatis, KeWang, and Nagendra Modadugu: The Ghost in the BrowserAnalysis of Web-based Malware, May 2007.13 David Barroso, ENISA Position Paper No. 3: Botnets – TheSilent Threat, November 2007,http://www.enisa.europa.eu/doc/pdf/deliverables/enisa_pp_botnets.pdf.14 Panda Security,http://www.pandasecurity.com/homeusers/media/pressreleases/viewnews?noticia=9077© 2010 ACADEMY PUBLISHER

Aims and ScopeCall for Papers and Special IssuesJournal of Emerging Technologies in Web Intelligence (JETWI, ISSN 1798-0461) is a peer reviewed and indexed international journal, aims atgathering the latest advances of various topics in web intelligence and reporting how organizations can gain competitive advantages by applying thedifferent emergent techniques in the real-world scenarios. Papers and studies which couple the intelligence techniques and theories with specific webtechnology problems are mainly targeted. Survey and tutorial articles that emphasize the research and application of web intelligence in a particulardomain are also welcomed. These areas include, but are not limited to, the following:• Web 3.0• Enterprise Mashup• Ambient Intelligence (AmI)• Situational Applications• Emerging Web-based Systems• Ambient Awareness• Ambient and Ubiquitous Learning• Ambient Assisted Living• Telepresence• Lifelong Integrated Learning• Smart Environments• Web 2.0 and Social intelligence• Context Aware Ubiquitous Computing• Intelligent Brokers and Mediators• Web Mining and Farming• Wisdom Web• Web Security• Web Information Filtering and Access Control Models• Web Services and Semantic Web• Human-Web Interaction• Web Technologies and Protocols• Web Agents and Agent-based Systems• Agent Self-organization, Learning, and AdaptationSpecial Issue Guidelines• Agent-based Knowledge Discovery• Agent-mediated Markets• Knowledge Grid and Grid intelligence• Knowledge Management, Networks, and Communities• Agent Infrastructure and Architecture• Agent-mediated Markets• Cooperative Problem Solving• Distributed Intelligence and Emergent Behavior• Information Ecology• Mediators and Middlewares• Granular Computing for the Web• Ontology Engineering• Personalization Techniques• Semantic Web• Web based Support Systems• Web based Information Retrieval Support Systems• Web Services, Services Discovery & Composition• Ubiquitous Imaging and Multimedia• Wearable, Wireless and Mobile e-interfacing• E-Applications• Cloud Computing• Web-Oriented ArchitectruesSpecial issues feature specifically aimed and targeted topics of interest contributed by authors responding to a particular Call for Papers or byinvitation, edited by guest editor(s). We encourage you to submit proposals for creating special issues in areas that are of interest to the Journal.Preference will be given to proposals that cover some unique aspect of the technology and ones that include subjects that are timely and useful to thereaders of the Journal. A Special Issue is typically made of 10 to 15 papers, with each paper 8 to 12 pages of length.The following information should be included as part of the proposal:• Proposed title for the Special Issue• Description of the topic area to be focused upon and justification• Review process for the selection and rejection of papers.• Name, contact, position, affiliation, and biography of the Guest Editor(s)• List of potential reviewers• Potential authors to the issue• Tentative time-table for the call for papers and reviewsIf a proposal is accepted, the guest editor will be responsible for:• Preparing the “Call for Papers” to be included on the Journal’s Web site.• Distribution of the Call for Papers broadly to various mailing lists and sites.• Getting submissions, arranging review process, making decisions, and carrying out all correspondence with the authors. Authors should beinformed the Instructions for Authors.• Providing us the completed and approved final versions of the papers formatted in the Journal’s style, together with all authors’ contactinformation.• Writing a one- or two-page introductory editorial to be published in the Special Issue.Special Issue for a Conference/WorkshopA special issue for a Conference/Workshop is usually released in association with the committee members of the Conference/Workshop like generalchairs and/or program chairs who are appointed as the Guest Editors of the Special Issue. Special Issue for a Conference/Workshop is typically made of10 to 15 papers, with each paper 8 to 12 pages of length.Guest Editors are involved in the following steps in guest-editing a Special Issue based on a Conference/Workshop:• Selecting a Title for the Special Issue, e.g. “Special Issue: Selected Best Papers of XYZ Conference”.• Sending us a formal “Letter of Intent” for the Special Issue.• Creating a “Call for Papers” for the Special Issue, posting it on the conference web site, and publicizing it to the conference attendees.Information about the Journal and Academy Publisher can be included in the Call for Papers.• Establishing criteria for paper selection/rejections. The papers can be nominated based on multiple criteria, e.g. rank in review process plus theevaluation from the Session Chairs and the feedback from the Conference attendees.• Selecting and inviting submissions, arranging review process, making decisions, and carrying out all correspondence with the authors. Authorsshould be informed the Author Instructions. Usually, the Proceedings manuscripts should be expanded and enhanced.• Providing us the completed and approved final versions of the papers formatted in the Journal’s style, together with all authors’ contactinformation.• Writing a one- or two-page introductory editorial to be published in the Special Issue.More information is available on the web site at http://www.academypublisher.com/jetwi/.

(Contents Continued from Back Cover)Webinar – Education through Digital CollaborationAnuradha Verma and Anoop SinghEfficient Visual CryptographySupriya A. KingerMultistage Interconnection Networks: a Transition from Electronic to OpticalRinkle Rani Aggarwal, Lakhwinder Kaur, and Himanshu AggarwalWeb Based Hindi to Punjabi Machine Translation SystemVishal Goyal and Gurpreet Singh LehalProtecting Data from the Cyber Theft – a Virulent DiseaseS. N. Panda and Vikram Mangla131137142148152