2005 Forum Home

Program

Keynote

Posters & Demos

GSRF

  > Org Committee

  > Prog Committee

  > Dates

  > For More Info

For Grad Students

Registration: Onsite

Maps & Directions

Past Forums

For More Info

A Light-weight, Event-based, Grid Infrastructure for Data-intensive Environments

Student: Chris Mattmann, USC/ISR

Collaborators: Daniel J. Crichton, JPL; Sam Malek, USC; Marija Mikic-Rakic, Google; Nels Beckman, USC

Advisor: Nenad Medvidović, USC/ISR

Abstract: In data-intensive environments, grid technologies have emerged at the forefront will commercial backing, and widespread use in academia, the open source community, and government projects. Each grid technology however has an underlying assumption that there are ample computing and hardware resources available with which to conduct science, information management, and solve highly complex problems. Most of these technologies consider the system's architecture, and its subsequent relationship to the implementation artifacts to be outside of their scope. Finally, each grid technology in some fashion relies on what could be considered a "heavyweight" middleware substrate to handle functions such as marshalling/unmarshalling of data,programming language abstraction, communication and the like.

To address these limitations, we have leveraged our experience in developing two distinct middleware solutions to arrive at a light-weight, grid-based middleware system for data-intensive environments called "GLIDE". GLIDE implements the grid services defined by the OODT middleware technology,developed at the Jet Propulsion Laboratory, and leverages the event-based,PRISM-MW middleware, developed at the University of Southern California. We have evaluated GLIDE in the context of a sample media sharing application deployed on a set of PDAs, and laptops, as well as in the context of a graduate-level software engineering course. GLIDE was found to be not only highly scalable, and have low memory footprint, but also to be a useful pedagogical tool in the context of the aforementioned graduate software engineering course titled "Software Engineering for Embedded Systems".

Our future work is centered on determining the appropriate smallest set of "grid services" which still provide ample utility in mobile, resource-constrained environments as it is our ultimate goal to "bring the grid to one's pocket".

Bio: Chris Mattmann is a Ph.D. student in the Computer Science Department at the University of Southern California, where he is a member of the Software Architecture Research group, advised by Dr. Nenad Medvidovic. Chris recently passed his screening exam (November 2004) and is currently formulating his dissertation topic, in the area of "Software Architecture for Data Intensive Systems". In parallel to his research at USC, Chris works full time at the Jet Propulsion Laboratory, as a Software Engineer in the Modeling and Data Management Systems section, where he is managed by Daniel J. Crichton.