Project Title

UDDI (Universal Description, Discovery and Integration) based Lookup Service for Grid Analysis Environment.

GAE (Grid Analysis Environment) and Lookup Service

Grid computing makes use of huge amount of resources by connecting large numbers of geographically dispersed individual computers, to facilitate the scientists for computation of large data sets. Grid Analysis Environment (GAE) provides a framework for physicists to analyze the results generated in Grid environment. Controlling the huge number of connected resources in such networks is nearly impossible using a centralized arbitrator. Lookup services are the main entry point to access the resources and services on the Grid. Locations of various services are not known in advance, the lookup service helps locate these services.

Aim & Objectives

The main aim of this project was to use UDDI for service publication and discovery. This new service should be compatible with existing lookup mechanism within GAE. But the centralized nature of UDDI becomes an obstacle when we want to achieve dynamism and scalability in our application. This dynamism and scalability is one of outmost requirement of GAE architecture. So we need to come up with a solution to avoid such hindrances.

Architecture

The replication among different nodes of uddi is achieved by using the MonaLisa. The MonaLisa framework provides a distributed monitoring service system using JINI/JAVA and WSDL/SOAP technologies. Each MonaLisa server acts as a dynamic service system and provides the functionality to be discovered and used by any other services or clients that require such information. MonaLisa allows its clients to publish different information to it repository which could be dynamically accessed from any other client. So using this mechanism, we achieved replication. Here are step wise procedures

• Publish

Jclarens publish its service description to uddi and also send this information to MonaLisa server. MonaLisa server stores all the information within its repository, this information is also time stamped. Jclarens keeps on publishing information with fixed interval. This step is necessary to remove stale service information from the lookup service.

• Discover

Service discovery can be directly done to the uddi registry using any uddi client compliant to UDDI version 2 specifications. We also provide uddi client API for accessing the service information from our UDDI registry. Service information is also downloaded by Jclarens periodically from the MonaLisa for ensuring replication of service information.