It's Cobbler TimeForever overlooked, IT management is now the focus of autonomics and next-generation technologyContinued from Page 1 Talking with Alan GanekIBM has clearly been a visible champion when it comes to setting the vision, doing the research and development, and bringing autonomic computing into IT systems, which will be an essential step toward IT infrastructure intelligence. In future columns, I hope to report on the views of a variety of IT systems providers and implementers about this important trend. In this column, I would like to close out with a brief interview with Alan Ganek, vice president of Autonomic Computing at IBM. Ganek joined IBM as a software engineer in 1978 and went on to win Outstanding Innovation Awards for his work on Enterprise Systems Architecture/370 and System/390 Parallel Sysplex Design, to manage the MVS and VM/XA products, and then to head technical strategy and operations for IBM's Research Division. In his career at IBM, he has also served as Programming Systems Director, Quality and Development Operations; Director of Solutions Development, IBM Telecommunications & Media Industry Unit; and several other managerial posts. IE: When most people think of automation, the point is to reduce cost. It seems like autonomic computing, however, has a broader goal. Ganek: "Autonomic" is really a metaphor for the human nervous system. You think about a baseball pitcher: He or she worries about where to pitch the ball for a given hitter, not about sweating and breathing. The body does those things automatically. Now, in IT, if you talk to CIOs and their staffs, you find out that people are bogged down, trying to keep systems running and determining where problems lie, rather than on new applications and getting what they want out of the whole system from a business point of view. IE: But isn't it possible that autonomics could make systems more rigid? Humans can only be humans; they can't turn themselves into elephants or cougars. Ganek: I should say first that autonomics are not about taking humans out of the picture. What we're looking for is a better balance between what the technology does and what people running the systems can do. To answer your question, right now, systems are so rigid that people are afraid to touch them. Tradition has been that you put a complex application into production and you tuned it until it ran well. And then you didn't touch it for as long as possible because you didn't want to mess it up. At some point, after a lot of testing, you might move to a new version of the application. That scenario is fine when the workloads you're dealing with are static. But in today's Internet world, you have unpredictable, explosive, widely changing workloads. To deal with this, most organizations will traditionally figure out where the peak is and then plan for the worst case. This is expensive. We have recently created new technology that allows IT to use a much smaller pool of resources, with the ability to dynamically and automatically pull in more from pools used for other applications. Ask someone how long it normally takes them to repurpose servers: often days or weeks. This is an example of how autonomics can dramatically increase flexibility and agility. With autonomic computing, we have three goals. Increasing ROI and reducing costs is certainly one major goal. Second is improving availability and reliability. The third is about time to value: You need to be able to deploy technology to solve a problem much more quickly. Today, complexity is a huge obstacle; working on projects with partners, such as the ASF with Cisco, is about dealing with the complexity so that organizations can introduce applications much more quickly and flexibly. IE: With problem determination — which looks like a major focus of the IBM/Cisco effort — many organizations face an information silo problem familiar to data warehouse and business intelligence application managers. Could the work with Cisco help IT integrate data and then apply analytics to figure out the most effective and strategic solution, given business objectives? Ganek: Yes. It used to be that companies had these lead gurus who understood everything that was going on. Now, with all the sophistication in the hardware, networks, and software, with different terminology and tools, you don't have that all-knowing guru. Having a format to define the semantics of the data that you capture will allow organizations to start doing intelligent correlation. Right now, even data about time is not captured uniformly. With a better structure, we can see the correlation between events or activities that start in the transaction manager and how they line up with activities in the database and network. "Sense and respond" is a central paradigm of autonomic computing. The sensitivity must be as granular as possible so that the data fed into analytical and predictive models gives the clearest picture possible. If you wait until you need to allocate more Web servers, for example, you're lost. You need to identify patterns, predict what you're going to need, and have that information available so that you can adjust as necessary. IE: What will CIOs be looking at five years from now? Ganek: Well, no one size fits all. CIOs who aggressively take advantage of autonomic computing will find that they can sustain much more growth and availability, but also manage it at a lower cost. Their role will finally move from keeping systems running to being a true partner in reaching business objectives. Autonomics will provide the flexibility to make that transition. David Stodder [dstodder@cmp.com] is editorial director of Intelligent Enterprise.
|
 |
Most Popular This Week
IE Weekly Newsletter
Subscribe to the newsletter
| |
| ||||||||||||||||||||||||||||||||
