E-merging tech: autonomic computing self-managing computing spread to partners, industry.
IBMs drive to make computing more self-managing is spreading beyond its product line to its business partners, software vendors and standards groups.
Autonomic computing is IBMs answer to what it sees as a core IT problem: how to make enterprise computing less complex. But despite its expertise, IBM officials know the company cant go it alone.
"Traditionally, weve approached this with system management tools," said Alan Ganek, vice president of autonomic computing at IBM, in Somers, N.Y. "That can help a lot, and its still a cornerstone of making progress, but you also need to attack the problem in the source. So we work on this across the company. But it really takes a broad-based and standards-based approach, so a multivendor solution is possible."
The term "autonomic" comes from the human bodys autonomic nervous system, which acts in the background to monitor its activities, Ganek said. IBM takes that analogy to the IT environment.
"Were working on standards, and were working with standards bodies," Ganek said. "We had an event at DeveloperWorks Live [in April in New Orleans] to focus on how we can bring this to developers. Whether those developers are application developers or product developers such as hardware or software developers, what were trying to provide are technologies people can get their hands on right now in tool-kit-type fashion and start to work."
Tripwire Inc. is one such company looking to work with IBM on autonomic computing, said Wyatt Starnes, CEO of the Portland, Ore., developer, which creates software for monitoring the integrity of enterprise data. Tripwire software senses unauthorized changes that cause system downtime. Starnes said the company has been working with IBM and is integrating its software "up into the [IBM] Tivoli platform. Were working to support some cross-platform standards and exploring the needs of the different [IBM] brands."
The bottom line, said Starnes, is "weve got to start treating IT as a system, not as a bunch of systems."
Marc Camm, chief operating officer at Adjoin Solutions Inc., in Boston, said his company is looking to tap into IBMs autonomic expertise to enhance its Web services systems management solution, Service Oriented Management and Monitoring Architecture. "The movement toward self-healing systems is something we feel we have to support in the data center."
Camm said Adjoins technology features "deep integration with Tivoli through the Tivoli management console." The Adjoin solution enables the system to discover what services are running and then self-manage and self-heal them as needed, Camm said.
Ganek said IBMs Tivoli system management division is synonymous with the autonomic push. "Virtually all of their products have been upgraded to introduce these kinds of intelligent algorithms" that promote self-monitoring and self-healing systems, he said.
The road to autonomic computing is a journey that entails a five-level adoption model that IBM uses for IT users, Ganek said.
The first level, Basic, is where users are not really managing their environment. Managed, the second level, is where enterprises start to use tools to track their environment, but IT staff still has to fix problems. At the third level, Predictive, the system does monitoring, correlates data and recommends actions. "We expect that over the next year and a half to two years, a much broader coverage of the system components will get up to that level," Ganek said.
At the fourth level, Adaptive, the system not only monitors, correlates and recommends action but also takes automated actions of its own. "We have some things that start to approach this area now," Ganek said. "For the large part, its going to take another three to five years to get broad coverage."
At the fifth level, Autonomic, "the system not only reacts in a very automated way, but the components and the actions are driven by business rules and business policies," Ganek said. "So you get an integration of your business and the IT capability that supports it. Now youre talking in the four-, five-, six-year-and-beyond time frame, though we are making good progress at the initial implementations of those things."