GRIDtoday Logo Hewlett-Packard

DAILY NEWS AND INFORMATION FOR THE GLOBAL GRID COMMUNITY /

( Previous Article )    ( Table of Contents )   

Special Features:

AUTONOMIC COMPUTING LOOMS ON THE HORIZON

Though the phrase "autonomic computing" sounds like the creation of a sci-fi writer's vivid imagination, it is in fact a reality -- and many vendors are developing such technology. IBM, Hewlett-Packard, and Sun Microsystems all have their individual autonomic computing efforts under way. Cassat Corp. and Stottler Henke Associates Inc., both specialty firms, are also undergoing these initiatives.

By allowing computers to configure, heal and optimize themselves, the industry hopes to help organizations that use grids and other highly distributed computing models. Such developments would allow reduced hands-on maintenance, in addition to higher reliability and resiliency rates.

Though full adoption of such technology is close to five years away, many systems management products are already incorporating elements of autonomic computing.

Government agencies have also begun to take notice of these developments. Energy Department Labs, NASA and DARPA have taken great interest, while the IRS and other non-scientific sectors are also looking into self managing technology for their business needs.

Proponents of the new technology strongly believe that the beginning steps, like clear security and incident response policies, need to be implemented now.

Because a lot of computing cost lies in the monitoring and tuning of IT components, autonomic computing would reduce cost dramatically by erasing the need for human intervention.

IBM, for instance, is seeking end-to-end self regulation of all their IT environments, including hardware, middleware and applications.

Autonomic computing, in addition to cost-cutting, could offer improved security as well. Automatic response to cyberattacks, like with DARPA's Self- Regenerative Systems program, is hoped enhance security capabilities.

Managing complexities among multiple computers is the most important autonomic computing ability. Because industry and government sectors are developing more and more distributed grid and cluster initiatives, self managing would help jobs flow smoothly through dispersed computing nodes.

The DOE is testing Agent-Based High Availability (ABHA) system smart job recovery software created by Stottler Henke, a software development firm, to help clusters and grids identify problems and recover automatically.

DOE's Lawrence Berkeley National Laboratory (LBNL) will use the software to analyze the data from a Brookhaven National Laboratory nuclear physics experiment. If a problem is encountered, the program will start the process again automatically.

ABHA will start deployment on a Linux/Intel cluster with several hundred nodes, but will become more available as demand for reliable clustering rises. The solution will help organizations to harness cluster power nationwide, without having to worry about access privileges to the site where the failure occurred.

In addition, scientists at NASA are looking to use autonomic computing for projects like the Mission Services Evolution Center, which will provide ground and flight systems with a common architecture. The center will become operational within the next year or two.

And the IRS is looking to incorporate the Tivoli tool suite for autonomic computing; it currently uses the Tivoli Enterprise Console and IBM Tivoli Monitoring 5.1. Sun Solaris servers already utilize some autonomic functions.

By defining processes, federal IT managers will be more prepared to automate them. The honing of IT management practices will help to identify relationships among components and operations.

( Top of Page )

( Previous Article )    ( Table of Contents )