GLIDE: A Grid-based Light-weight Infrastructure for Data-intensive Environments


GLIDE provides a suite of grid services (e.g. resource discovery, resource description and search and retrieval) layered on top of an event-based middleware for resource-constrained devices. In effect, it represents the marriage of the Object Oriented Data Technology (OODT) data grid middleware, developed at JPL, and the Prism-MW middleware, developed at USC.

Programmers use GLIDE's Product Server component to abstract away interfaces to heterogeneous underlying resources (e.g. websites, file systems and storage systems). Profile Server components describe the type and location of resources in a GLIDE system deployment. Query Servers are GLIDE components which are used to integrate and aggregate resources that satisfy a particular query, and deliver the aggregrate resources back to the requestor.

Since GLIDE was developed using the Prism middleware, programmers and designers can design their grid systems at the architectural level. Grid services are built in the Component (units of computation), Connector (interaction facilities) and Configuration (topology of the system) architectural style. The design of the grid system is carried through into the implementation artifacts (i.e. the code) via Prism's architectural facilities. GLIDE systems rely on Prism for underlying middleware facilities such as programming language abstraction, messaging, event marshalling and unmarshalling and transformation (such as encryption).


The latest version of GLIDE is version 1.1.8. Since GLIDE is implemented using the Java programming language, it runs on convential PC desktops with a 1.1.8 version Java Virtual Machine (JVM) running LINUX/UNIX, or Windows. Additionally, GLIDE runs on the Jeode Runtime for Compaq IPAQ PDAs.

You can download the zipped file here.

Installation Requirements

To install and run GLIDE, you'll need the following external components:

Ensure that the necessary jar files/components are identified in your CLASSPATH environment variable before trying to run GLIDE.


At the current time, we have compiled an installation guide/user's manual for GLIDE, and it's available here. We're working on further documentation. Additionally, you can find the javadocs for GLIDE here.


Listed below are the relevant publications on GLIDE:

C. Mattmann, N. Medvidovic, P. Ramirez and V. Jakobac. Unlocking the Grid. To appear in Proceedings of the 8th ACM SIGSOFT International Symposium on Component-based Software Engineering (CBSE8), St. Louis, Missouri, May 14th-15th, 2005.

C. Mattmann, S. Malek, N. Beckman, M. Mikic-Rakic, N. Medvidovic and D. Crichton. GLIDE: A Grid-based, Lightweight, Infrastructure for Data-intensive Environments. In Proceedings of the European Grid Conference (EGC2005), LNCS 3470, P. M. A. Sloot, A. G. Hoekstra, T. Priol, A. Reinfield, M. Bubak (Eds), Springer Verlag, Amsterdam, The Netherlands, February 14-16, 2005.