Standards Need Accord
For it to get to the next level, both in terms of technology and in the eyes of business managers, standards must be developed with input from all major stakeholders, and they must be widely supported and fiercely protected. If this does not happen, technologies that promise to save companies time and money and build competitive advantage simply wont take root.
Take autonomic computing, an umbrella term for IT products that configure, heal, optimize and protect. Is it coming soon to an IT department near you?
IBMs vice president of autonomic computing, Alan Ganek, told me the companys autonomic computing initiative is making its way into hundreds of applications. Extrapolate this over the next several years, and we could see IT components ranging from servers, routers and wireless access points to databases and operating systems acting autonomously.
Autonomic computing is more than just the next buzz term or marketing hooey. Indeed, this coming transformation should be embraced for two important reasons.
First, vendors making autonomic computing products have real experience and research behind themnot just smoke and mirrors. IBM, Computer Associates, Hewlett-Packard and others have had nearly 10 years to get enterprise management right. These vendors learned hard lessons from their enterprise management platforms stumblesknowledge that will likely take the form of autonomic computing.
Second, autonomic computing, leveraging products that these vendors have honed for many years, can have a direct, significant impact on a companys bottom line: The speed with which autonomic systems can allocate and reallocate computing resources based on business rules, for example, is immeasurably faster than people sitting around in conference rooms making action plans.
But autonomic computing faces huge challengesmost notably, a lack of consensus on standards. Although many standards bodies and vendor consortia are working to bring order to chaos, it will likely be a cold day in hell before these groups reach anything close to consensus. And consensus is critical for autonomic computings success.
Unlike telephone companies or automobile manufacturers, which have been subjected to various levels of government mandates for interoperability (think seat belts, air bags and impact tests), IT has been left entirely subject to market forces.
The result is an unsurprising mess. Combine "Might makes right" (Microsoft) with "fear, uncertainty and doubt" (IBM) with "Its not my fault" (any software license agreement), and cooperation in the name of reducing customer management costs goes out the window.
IBM is trying to change this because its becoming clear that for IT resources to continue to drive productivity increases at levels customers have come to expect, management and maintenance costs have to be radically reduced. This is good news for enterprise consumers of IT products, but it means that CIOs and CTOs need to get serious about helping standards bodies enforce rules.
For example, IBM is trying to inject life into a languishing initiative called ARM, or Application Response Measurement. When I first started writing about IT back in the late 1990s, ARM sounded like a good idea. Its been in the morgue ever since because application vendors saw performance reporting as almost an act of suicide (especially when most performance problems had little to do with the application but rather were caused by any number or combination of network and system components that the application could do nothing about). Application makers knew the messenger would get shot, so, for the most part, they stayed away from equipping their wares with open performance monitors.
IT managers should push for resources to participate in various standards or specification groups that are working on the fundamental guidelines that will help make some form of autonomic computing a reality. The Global Grid Forum, OASIS, the DMTF and several Web services groups are worth participating in so that the voice of IT consumers is heard at this crucial, formative stage of autonomic computing.
Senior Analyst Cameron Sturdevant can be contacted at cameron_sturdevant@ ziffdavis.com.