IBM has announced new software that will address a major source of inefficiency in computer networks and could drive more mainstream commercial implementations of grid computing - the ability to take untapped resources across an enterprise and make them available where and when they’re needed, resulting in a single, virtual system.
“In the increasingly complex data centres in the region, clusters of servers that run business applications often do a poor job of juggling unpredictable workload. One server may sit idle, while another is overloaded,” explained Bashar Kilani, manager of IBM Software Group for the Middle East, Egypt and Pakistan. “This leads to a Catch-22 situation where companies often plan for the highest possible spikes in demand, then watch as all those servers operate well under capacity most of the time. Organisations in the Middle East are looking for better ways to use the hardware and software they have already purchased.”
IBM has developed a first-of-its-kind “traffic cop-like” software product that automatically and intelligently monitors application workload and routes traffic to one server or another according to its workload at a given time. It allows a cluster consisting of many servers - from dozens to hundreds - to operate as a single environment that automatically adapts to sudden changes, much as the electrical grid works. This can improve network performance, for example, eliminating some of the missed server connections that can be a mystery to Web users, and allow companies to get more value out of their existing server resources.
The technology, developed by IBM Research and the company’s software development teams over the past two years, is available in the IBM WebSphere application server. A top analyst firm recently named WebSphere as the best-selling Web application platform.
The “traffic cop” is easily installed by an IT administrator with a simple point and click as part of the WebSphere software setup.
The new technology is the latest example of IBM’s work to take grid computing into the commercial mainstream. IBM is offering technology and bringing together business partners, developers, system integrators and IT vendors of all sizes to form a standards-based grid ecosystem that will accelerate adoption in the enterprise.
In grid computing, all of the disparate computers and systems in an organisation or among organisations become one large, integrated computing system. That single system can then be turned loose on problems and processes too large and intensive for any single computer to easily handle in an efficient manner. Under non-grid circumstances, mainframes might lie idle 40 percent of the time. Unix servers are sometimes “serving” less than 10 per cent of the time. Many business PCs may be under-utilised as much as 95 per cent of a typical day.
Future versions of IBM WebSphere will extend the “traffic cop” capability to disparate parts of a company and automatically coordinate multiple clusters of servers running various business applications, rather than just single clusters. Other new features in WebSphere that help customers improve network performance are:
1) WebSphere Performance Advisor: This simplifies the IT administrator’s job by using live data collected from a running system to analyse changes and recommend actions to improve application performance. It advises IT managers how to set up the system to handle different levels of network traffic and also specific parameters. While this feature helps customers anticipate changes in their environment today, future improvements will provide more advanced autonomic features.
2) Automatic Backup Clusters: With WebSphere, customers can automatically configure their system to set up a backup cluster of servers in case the primary cluster fails, without having to write any code.
With V5.0.2, WebSphere also broadens its support for a multi-platform open standards-based IT environment by adding Microsoft’s Windows 2003 server to the 25 platforms that WebSphere already supports.
