RSS
Twitter
Self-Healing Computers Still A Long Way Off
from the heal-just-a-tiny-part-of-thyself dept
It's been a few years now since IBM made a big announcement about plans for self-healing "autonomic" computing systems, and while there have been some advancements in the space, they've been slow. Business Week now has the latest report, which tries to sound up-beat, but it still seems like the advancements are still coming along quite slowly. Given the challenge involved, perhaps it's not that surprising -- but so far it's falling well short of the picture companies were painting a few years ago about computers that automatically would spot a crash or a problem, isolate it and solve it without bothering a human. So far, it looks like only the very basics have been accomplished -- adding a bit of smarts to various systems to better understand and route issues and maybe trying to give a bit of an early warning if something is really going wrong. It's good to see at least some advancements, but it sounds like the big breakthroughs are still going to take plenty of time.
Reader Comments (rss)
(Flattened / Threaded)
Back in the 1980s
they had already built microchips that could rewire themselves with tiny switches. I guess it remains uneconomical.
(reply to this comment) (link to this comment)
Re: Back in the 1980s
This is a classic touring problem. It's unsolvable.
(reply to this comment) (link to this comment)
Touring problem
This sounds like a classic touring problem, unsolvable.
(reply to this comment) (link to this comment)
No Subject Given
Didn't Sun already start paving the way with self healing in Solaris 10?
(reply to this comment) (link to this comment)
Re: Touring problem
Are you talking about the NP-completeness problem, or the halting problem of Turing machines?
(reply to this comment) (link to this comment)
No Subject Given
What if the self healing dingus fails?
(reply to this comment) (link to this comment)
Add Your Comment