KIG - Team Interests

Written by Super User. Posted in Research

Team  interests

Knowledge and ontology supported metadata management for the Grid

According to Mario Campolargo [CAM06], obtaining knowledge from data and knowledge management is intended to be one of the main focuses of 7FP. Our experience in this field has been obtained during the Pellucid project, in which we proposed and implemented organizational memory based on ontological approach  and strategies for knowledge reuse with the help of software agents [LAC04]-[SLO07]. For the purpose of semi-automatic workflow construction for grid applications (in K-Wf Grid) the distributed grid organizational memory subsystem has been developed ( , that contains different kind of ontologies (generic and application domain specific), most notably: application, resource, data, service and workflow ontologies ([KRY06-1]-[KRY07]). Algorithms for knowledge extension (based on ontology interoperability and alignment) are being developed [SLO05]-[SLO06-2].
Our future work in this field is planned toward further development of ontology evolution methods and interoperability for collaboration of dynamically created virtual organizations.

Cited literature for Knowledge and ontology supported metadata management for the Grid

Grid Virtual Organizations

In the relatively short history of the Grid computing, Virtual Organizations have been understood in a variety of ways, from low level account management support to high level business process modeling. In any case, Virtual Organizations are about sharing private resources of collaborating partners in order to achieve a goal which otherwise would be much more difficult to achieve. In order to allow heterogeneous partners to form and take part in such virtual collaborations, their resources that they are willing to share must be properly virtualized. These include data, services and domain knowledge. Also security and trust issues must be taken into consideration especially when dealing with business entities.
Recently, a Grid Virtual Organization Semantic Framework is being developed (within the scope of GREDIA project), which will enable both business and societal partners to collaborate together using their resources to challenge emerging market opportunities [KRY06-4] with the focus on multimedia and banking segments.  Extension of this functionality of the framework will allow data delegation for remote (grid) processing with real-time processing requirements. This Virtual Organization environment is being developed (in project [DUT06-1]) for an interactive, real-time High Energy Physics application, that requires dispatching processing of high frequency events generated in ATLAS detector to remote Grid sites for processing [SKI06-1], [PIE07-1].
Cited literature for Grid Virtual Organizations


Semantically supported Grid interoperability

Interoperability between two or more Grid environments means that all crucial components of each Grid middleware can communicate in order to transparently execute users' requests using resources available through the middleware of the interoperating infrastructures. Thus one of the key issues and obstacles at present is how to refer to heterogeneous resources managed by different Grid middleware systems. We believe that semantics, for instance ontologies, used to represent metadata about these resources will foster much more flexible integration of these Grid middleware layers in the most transparent way. Higher level metadata of resources in the Grid is in fact crucial when trying to create and manage Virtual Organizations spanning over multiple heterogeneous Grid middleware systems, especially when other issues such as cultural heterogeneity take place [KRY06-5].
We have already made some analysis of possibilities of realizing this vision. We've proposed a Grid Abstraction Layer for providing interoperability between gLite and Grid Operating System which addressed such Grid use cases as job submission and data management [KRY06-6].
Cited literature for Semantically supported Grid interoperability


Data management in the Grid

Data are the most important assets in every activity. Its availability and efficient access are crucial problems of contemporary science and business. One of the achievements in this field is Component-Expert technology that has been applied for proper selection of Storage Elements in the Grid (in Crossgrid, [DUT02]-[DUT04] as well as for service selection in semi-automatic construction of workflows (in K-Wf Grid, [DUT05-1]. To support this selection, another activity in this field on storage modelling and simulation of data access time for different storage systems (like HSM and secondary storage, [NIK02-1]-[NIK04-2]) has been initiated.  Algorithms for virtualization of the storage systems and for automatic data replication in the grid environment have also been developed ([NIK04-2]-[SLO06]).
At present our focus is on development of WS and WSRF type services for estimating time of data access in mass storage systems as well as on ontology of secondary and tertiary storage systems, with the intension to be interoperable with CIM ontology. For that purpose algorithms for translation of CIM/MOF to OWL ontology are being developed [MAJ06].
Cited literature for Data management in the Grid