Computing Seminar

Knowledge Grids

Carl Kesselman, USC/Information Sciences Institute

Abstract

We are at the brink of an information technology revolution. Web browsers are now nearly as common place as telephones, and one can no longer imagine a world in which documents of various types are not shared across broad communities of users. Future versions of the Web promise to be even more powerful as the web shifts from describing document structure to describing the semantics, or content of the documents. Meanwhile, the past several years have seen the development of technology known as the Grid, whose goal is to allow ubiquitous access to a wide range of different types of resources: computers, software, storage, telescopes, microscopes, and virtually anything else that can be attached to a network. By breaking down geographic and organizational barriers to resource access and use, Grid technology promises to fundamentally change the way computers are used as well as the way people use information technology to collaborate with one another. As important as these two technologies are by themselves, taken together, these technologies have even greater potential: generalizing the semantic web by allowing any type of resource, not just documents, to be discovered and shared, allowing detailed semantics to guide discovery, selection, configuration and use of resources on the Grid.

In this talk, I will explore the synergistic relationship between the Semantic Web and the Grid. I will introduce Grid technology with a focus on recent developments in service oriented approaches to Grid design, specifically the Open Grid Services Architecture. Grid services build on Web services standards, extending them by defining basic semantics for what it means to be a service, and creates a model for transient, statefull service instances. I will then discuss the value that Semantic Web technology such as ontologies and inference can bring to the Grid, and how the Grid can generalize the applications of the Semantic web. I will conclude by describing new classes of applications that are possible with such an integrated infrastructure.

About the speaker: Carl Kesselman is one of the co-founders of the Globus project (http://www.globus.org). He is the director of the Center for Grid Technologies at the University of Southern California's Information Sciences Institute, and chief software architect for the National Project for Advanced Computational Infrastructure.