JERA

Core features:

 

Distributed Web Caching system.

 

Distributed Index

 

Cache  management

 

 

 

 

 

 

 

 

 

 

The immense growth of the number of Web Servers and the pressure on the existing infrastructure forces the Internet community to investigate new models for Web exchange. With the increasing dependency on the supply of Digital Information, the existing ad hoc solutions that have been in use since the early days of the World Wide Web have to be reconsidered, and new, reliable solutions will have to emerge.

The JERA Project Goals

The JERA project focuses on the realization of a fault tolerant Distributed Web Service. Several aspects taken into account are described below:

Object Caching

To improve service times for users, frequently requested information from remote sites are kept in local storage. A possible implementation of this caching service is a hierarchy of object caches, possibly geographically distributed over the service region, where each user is connected to a node at a certain level in this hierarchy.

Data Base

Web applications are build more and more on high performance database systems. To allow for a good scalability the object oriented Matisse Database System is migrated to an MPP environment to enable a performance evaluation of the database management system on different hardware architectures, the port of the Matisse Server database management system is being implemented on a MPP architecture.

Fault Tolerance

When providing information that is essential for a clients production process, the continuous availability of the information must be guaranteed. Providing a reliable service with high availability requires a server that has no single point of failure so that no one part in itself is responsible for the service to operate. Incorporating fault tolerance into the server design enables the server to operate correctly, even in the case of soft or hardware failure.

Scalability

In order to keep up with the expected growth of the World Wide Web, the server needs to be adjustable to an increasing number of requests. The design should provide scalability on every level of the service, for all the server components. This involves investigations of different machine architecture used for different Web Server components, including a cluster of workstations, SMP and MPP machines.

Caption describing picture or graphic.

< Return to Project List

Workflow & Grid Middleware Group (WGM)