What the Grid can do for the Semantic Web

David De Roure

This is the abstract for the invited presentation 'What the Grid can do for the Semantic Web' at the Dagstuhl Seminar, July 2005.

One of the reasons that people turn to the Grid is for large scale computation and data processing, and there are Semantic Web applications which stand to benefit from this, including knowledge discovery and large reasoning tasks. However there are several additional ways in which the Semantic Web can benefit from the Grid and the Grid community.

• The Grid provides a robust, deployed distributed infrastructure for use and indeed for exploration - richer than Web Services at this time.
• Semantic Grid applications developed in recent years provide case studies which can inform Semantic Web research - for example the CombeChem e-Science project has over 80 million RDF triples across multiple stores and represents a substantial Semantic Web deployment, while other projects have emphasised issues of real time processing and interaction.
• Significantly, the Grid community has processes in place to support community interaction and development of standards, providing a mechanism to achieve the degree of interoperability which is needed for uptake of Semantic Web.
• It also provides an enthusiastic developer community actively seeking solutions to problems in Grid middleware and applications and willing to try new solutions.
• Additionally the Grid provides a real context for deploying Semantic Web Services, particulary to support Virtual Organisations, and Semantic Grid researchers are beginning to explore knowledge services supporting Grid services and vice versa.

Finally, a series of challenges for Semantic Web have emerged, including the engineering and architecture of large triplestores, working with incomplete or inconsistent descriptions, the relationship between Grid Services and agents, and autonomic aspects. In particular, Virtual Organisations are an important focus at this time, also drawing on aspects of agency.