Autonomic computing: freedom or threat?
Glenn Fink, Deborah A. Frincke · 2007
No longer is the question whether autonomic computing will gain general acceptance but when. Experts like Alva Couch expect autonomic computing to be widely used within 10 years [1]. When it does become mainstream, how will autonomics change system administration and corporations, and will the change be for better or worse? The answer depends on how well we anticipate the limitations of what autonomic systems are suited to do, whether we can collectively address the vulnerabilities of autonomic approaches as we draw upon the advantages, and whether administrators, companies, partners, and users are prepared for the transition. This article presents some design considerations to address the first two issues and some suggested survival techniques for the third. I. What is Driving Autonomic Computing? Computing systems used to have reasonably well-defined borders, both geographically and logically. Today, corporate, educational, and even government computer-based systems coexist in an open mesh of overlapping infrastructures. The complexity of these networks and of the systems that comprise them have given rise to the need for more automation in their configuration and management. Companies rely on a growing number