1 - Acta Technica Corviniensis

traditional MANET routing protocol cannot be used insparse MANETs. A key challenge is to find a route thatcan provide good delivery performance and low endto‐enddelay in a disconnected graph where nodesmay move freely.To overcome this issue, node mobility is exploited tophysically carry messages between disconnected partsof network. The scheme that exploits the nodemobility, referred to as mobility assisted routing thatemploys the store‐carry‐and‐forward model is used.Mobility assisted routing consists of each nodeindependently making forwarding decisions that takeplace when two nodes meet.In VANET, when few vehicles are equipped withwireless transceivers, network will be sparse; delaytolerant routing algorithms are needed. The proposedMotion Vector Algorithm (MOVE) [8] for V2R VANETconsiders sparse network where prior prediction mustbe made for rare opportunistic routing. It is assumedthat every node has knowledge of its own positionand heading, where destination is a fixed globallyknown location. From this current vehicular nodefinds closest distance between vehicle and messagedestination along its trajectory. Current vehicularnode periodically sends HELLO message. Neighbouringnodes sends RESPONSE message to make itself knownto current vehicular node. Given the direction ofwhere neighbouring node is heading; current nodedetermines the shortest distance to destination alongthe trajectory of neighbouring node. The current nodethen makes decision to forward the message whiledetermining the each vehicle’s current distance fromdestination. This algorithm where data delivery rate ishigher for sparse network, compared to greedy,position based routing and uses less system bufferspace. With resulted performance evaluation, authorshave noted that if routes are consistent and uniform,greedy position based routing performs better thanMOVE.In line with MOVE algorithm another algorithm calledScalable Knowledge based Vehicular Routing (SKVR)[1], also makes the usage of the predictable routes andvehicle schedules. It divides the network in interdomainand intra‐domain. In inter‐domain routingsource and destination belong to different routeswhereas in intra‐domain source and destinationbelong to same route. In inter‐domain algorithm,message is forwarded to a vehicle travelling indestination domain and once destination domain isreached intra‐domain message delivery procedure willbe followed. In intra‐domain messages are sent inforward or reverse directions, depending on theentires of contact list. If the sending vehicle contactlist does not contain any vehicle in the destination’sdomain, then messages are delivered to the otherACTA TECHNICA CORVINIENSIS – Bulletin of Engineeringvehicles in contact list. When vehicles along the sameroute encounter one another, a node carrying amessage must decide whether to continue bufferingthe message, or to forward it, based on the directioninformation of the vehicle.Using strategy called ‘carry‐and‐forward’ VehicleAssisted Data Delivery (VADD) [17] algorithm allowspackets to be carried by vehicles in sparse networkand eventually relaying it to appropriate node when itenters in broadcasting range. Each node in VADDknows its own position and also requires externalstreet map that includes traffic statistics. Selection ofthe candidate node, to which message need to beforwarded, is encountered through different selectioncriteria. However such criteria are either not scalableor consumes more bandwidth through duplication ofpackets. Authors have observed while using VADD,network becomes unstable as vehicle densitydecreased, because optimal paths were not availableand because algorithm relies upon probabilistic trafficdensity information.Unlike VADD, Static Node Assisted Adaptive Vehicularrouting (SADV) [6] where static node has capability tostore a message until it can forward the message to anode travelling on the optimal path. Algorithm alsodynamically adapts to varying traffic densities innetwork, so that every node can measure the amountof time required to deliver message. However like any‘store‐and‐forward’ this algorithm requires theefficient buffer management. By using ‘Least DelayIncrease’ strategy, where static node checks whichpaths are currently available and eliminates packetswhich will not significantly increase their deliverydelay.Routing called Geographical Opportunistic (GeOpps)[9] routing in delay tolerant network is usingopportunistic routing with carry‐and‐forwardapproach to route messages. Algorithm assumes thatvehicle is using GPS and Navigation system that helpsto route and locate static road site unit.D. Quality of Service (QoS)QoS routing strategy is not followed by any traditionalMANET routing protocols. However there are researchattempt to integrate such strategies within MANETrouting protocols.Multi‐hop Routing Protocols for Urban VANET (MURU)[13], estimates quality factors of a route based onvehicle position, speed and trajectories. Based on thisquality factors MURU introduces new metric called‘Expected Disconnection Degree’ (EDD). Hence MURUnodes need to know its own position and haveexternal street map including presence of efficientlocation service. This new metric value considered tobe low as EDD, is an estimation of probability thatdetermines the breakability of route during given time1042012. Fascicule 3 [July–September]