The Relational Grid Resource Project

Relational Grid Resources (RGR) Project

The Relational Grid Resource project explores the representation of grid resource data. Grid information includes information and attributes about resources in a computational or data grid, including compute servers, clusters, storage resources, people, groups of people, etc. We posit that the needs of the computations that run on a grid impose a unique set of demands on a resource repository that are not being met by the current hierarchical or attribute-value approaches to directory services. Specifically, user-driven needs necessitate support for complex queries over the resource repository and the timeliness needs require policies about update rates to the data contained in the repository. Our approach explores relational database-based technology for management of grid resource information.

This work is being done in conjuction with the Unified Relational Grid Information Services project under Peter Dinda at Northwestern.

People

Beth Plale, faculty
Craig Jacobs, Professional Masters student
Scott Jensen, PhD student
Deepti Kodeboyina, Masters student
Ying Liu, PhD student
Charlie Moad, Professional Masters student

IU RGRbench Benchmark/Workload

RGRbench is developing a database benchmark/workload for purposes of better understanding relative strengths of different platforms for grid information servers. A grid information server is an essential component of the distributed grid middleware as it is a repository of resource information about entities on the Grid.

The benchmark/workload is a set of queries and 'scenarios', or short-duration workloads, developed from a common representation for resource data that is neutral to a particular data model.

We apply the benchmark to three platforms that could viably host a grid information server: one platform is a relational data model and uses SQL as the query language. The second and third are hierarchical data models: an XML system that uses XPath, and MDS2 which uses LDAP as its query language. Included also are scripts that operate on a relational database through the UK National e-Science Institue's OGSA-DAI.

Download

December 8, 2003: Version 1.0, a gzipped tarfile.

RGRbench and OGSA-DAI

Click here for information on our work with the UK e-Science OGSA-DAI data access and interface framework to OGSA grid services.

RGRstream

RGRstream explores data freshness and provenance in Grid information services/registries by employing continuous queries over streams for dynamic data and traditional queries over a database for data of a more static nature. This project leverages the dQUOB system.

Publications

Beth Plale, Craig Jacobs, Scott Jenson, Ying Liu, Charlie Moad, Rupali Parab, and Prajakta Vaidya, Understanding Grid Resource Information Management through a Synthetic Database Benchmark/Workload conference submission, under review, November 2003.
Beth Plale, Craig Jacobs, Ying Liu, Charlie Moad, Rupali Parab, and Prajakta Vaidya, Benchmark Details of Synthetic Database Benchmark/Workload for Grid Resource Information Technical Report TR-583, August 2003.
Beth Plale, Peter Dinda, and Gregor von Laszewski, Key Concepts and Services of a Grid Information Service, ISCA 15th International Parallel and Distributed Computing Systems (PDCS), September 2002, pdf.

Acknowledgement

This material is based upon work supported in part by the National Science Foundation under Grant No. ACI-0128390. Any opinions, findings and conclusions or recomendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF).

Last updated December 2003