14<strong>CoreGRID</strong>InstitutesThe general goal is t<strong>of</strong>urther integrate datamanagement andknowledge discoverysolutions with Gridtechnologies.Domenico TaliaLeader <strong>of</strong> the Knowledge and DataManagement Institute, Università dellaCalabriaKNOWLEDGE ANDDATA MANAGEMENT▼Grids are changing their role, moving from acomputation and data management plat<strong>for</strong>m to apervasive in<strong>for</strong>mation and knowledge managementinfrastructure. This trend requires new models,services and technologies <strong>for</strong> enabling Gridcomputing systems to manage distributed dataand knowledge, enabling complex applicationsaccording to the SOKU model.The INSTITUTE ON KNOWLEDGE AND DATAMANAGEMENT (KDM) joins together thirteeninstitutions from eight European countries andinvolves more than 50 senior researchers and PhDstudents. The general goal <strong>of</strong> this Institute is t<strong>of</strong>urther integrate data management and knowledgediscovery solutions with Grid technologies <strong>for</strong>providing data- and knowledge-intensive Grids.The Institute provides a collaborativesetting <strong>for</strong> European researchteams working on: distributedstorage management on Grids;knowledge techniques and tools<strong>for</strong> data-intensive applications;security and trust mechanisms <strong>for</strong>storage and data; and integration<strong>of</strong> data and computation Grids within<strong>for</strong>mation and knowledge Grids.The goal <strong>of</strong> the second year hasbeen to consolidate and expandthe joint activity <strong>of</strong> research groups,thus promoting larger leadingteams and supporting ef<strong>for</strong>tstowards standard models, services,middleware and solutions.During the third year, the Institutemembers (CETIC, FORTH, CNR-ICAR,INFN, CNR-ISTI, PSNC, STFC-RAL,SZTAKI, Universidade Nova de Lisboa,University <strong>of</strong> Calabria, University <strong>of</strong>Cyprus, University <strong>of</strong> Manchester,and University <strong>of</strong> Newcastle)worked on three main tasks:1 DISTRIBUTED STORAGEMANAGEMENTProviding infrastructures, techniquesand policies <strong>for</strong> managing distributedstorage resources in the Grid.2 INFORMATION AND KNOWLEDGEMANAGEMENTDeveloping metadata, semanticrepresentation and protocols <strong>for</strong>Grid service discovery, in<strong>for</strong>mationmanagement and design <strong>of</strong>knowledge-oriented Grid services.3 DATA MINING AND KNOWLEDGEDISCOVERYDesigning Grid services <strong>for</strong> distributeddata mining and knowledge discoveryon Grids and P2P systems.In all those areas in 2007 the partnersjointly produced scientific resultsimplemented in s<strong>of</strong>tware prototypesand published in scientific journalsand conference proceedings.
➦Data Mining and Knowledge DiscoveryKDM➦Distributed Storage Management15<strong>Annual</strong>Report2007In<strong>for</strong>mation and Knowledge Management➦Additionally, security issues havebeen covered with a particular focuson security requirements and models<strong>for</strong> storage and data management.ObjectivesThe main objective <strong>of</strong> the KDM Instituteis to address issues like representation,discovery and integration <strong>of</strong> data andknowledge in Grids and P2P systems.The Institute’s approach is a verticalapproach that encompasses all layersinvolved with knowledge managementin Grids: storage managementat the systems-level, in<strong>for</strong>mationand knowledge management, andknowledge discovery. At the sametime, single KDM solutions that bringbenefits to the Grid community havebeen designed and developed.A key KDM objective is to designand develop common solutions <strong>for</strong>data management and knowledgediscovery and management on Grids.This promotes the wide diffusionand use <strong>of</strong> knowledge-based Gridservices <strong>for</strong> the Semantic Grid andthe Knowledge Grid. To this end,the partners <strong>of</strong> the KDM Institutefocus on the problems <strong>of</strong> providingcommodity-based connectivity amongheterogeneous distributed storagedevices, management automation<strong>of</strong> administration tasks traditionallyhandled manually, and storagevirtualisation <strong>for</strong> serving well-definedrequirements from multiple users.Another challenge we investigatedis the need to develop autonomoussecurity mechanisms <strong>for</strong> decentraliseddata management involving sel<strong>for</strong>ganisingbehaviour. Finally, significantresults have been achieved in topicssuch as data federation, integrationand querying, dynamic workflowexecution, Grid-aware knowledgediscovery, and P2P data management.As a result <strong>of</strong> the joint activities, newresearch results, s<strong>of</strong>tware prototypesand system solutions have beendeveloped in such areas as storagemanagement, distributed dataintegration, reliable dynamic queryprocessing, storage security models,knowledge discovery services, Gridontology, and P2P data management.We organised Institute meetings andworkshops jointly with other <strong>CoreGRID</strong>Institutes, exchanged visits, set up jointfellowships, and produced two books,as well as tens <strong>of</strong> publications andtechnical <strong>report</strong>s. We organised jointresearch activities <strong>for</strong> research groupsworking on the main KDM topics. Jointactivities are carried out with other<strong>CoreGRID</strong> Institutes on scheduling <strong>for</strong>data Grids, self-configuring storage,scalable Grid services, resourcediscovery, and P2P Grids. We alsocontributed to the EC TechnicalGroup 5 on Data Management.Industryadded-valueThe results <strong>of</strong> the research activities<strong>of</strong> the KDM Institute can be exploitedby European industries workingin the area <strong>of</strong> data managementand knowledge services. Researchactivities in the area <strong>of</strong> data Grids andknowledge Grids are being pursuedby many industries such as HITACHI,Micros<strong>of</strong>t, IBM, SAP, ORACLE, andSIEMENS. This is testimony to theindustrial relevance <strong>of</strong> KDM topics.KDM Institute partners started newnational and European joint projectsin KDM research areas. <strong>CoreGRID</strong>partners submitted together withindustrial partners project proposals<strong>for</strong> building knowledge and datamanagement solutions made largelyusable through Grid services. Securityand service level agreement issues arealso investigated. Institute partnersworked with IAB member companieson proposing <strong>CoreGRID</strong> fellowshipsin KDM areas <strong>of</strong> industrial interest.Challenges andperspectivesDeveloping knowledge-basedapplications that exploit Grid featuresto achieve high per<strong>for</strong>mance andhigh availability is an area <strong>of</strong> activeresearch in Europe and overseas.Industry is also very active inexploring and adopting new solutions.<strong>CoreGRID</strong> KDM researchers willwork towards providing a variety <strong>of</strong>models, architectures, prototypes,and services <strong>of</strong>fering differenttechnological solutions to currentproblems faced by Grid applications.Future research activities will focuson core technologies needed <strong>for</strong>implementing the SOKU model,including ontologies, data miningand knowledge discovery, anddata management, also takinginto account SLAs. Partners <strong>of</strong>the KDM Institute will proceedbeyond the funded period and willcontinue their joint activities.