This week, I begin a series that focuses on an interesting but little noted technology direction that is likely have a profound effect on storage over the next few years. In fact, there is a better than ever chance that this technology concept will have a significant impact on most other aspects of computing systems as well.
It has been noted here and elsewhere that as IT managers we constantly find ourselves in a reactive mode in respect to many of the events that impact our systems. Despite our best planning, things break and we have to fix them. Systems degrade over time, and we spend a lot of cycles bringing them back up to acceptable levels of service. We make mistakes, and have to reconfigure things. We hot-swap drives, or fans, or power supplies in our RAID systems.
So we can be forgiven if every now and then we play the "what if" game. What if, for example, all the systems our staffers spend so much time coddling were able to care for themselves?
What if they configured themselves, and planned against future eventualities? And what if when they made a mistake in this planning, they learned from that error and built the results of that learned experience into their next series of decisions?
"Autonomic computing" is the name generally used to describe many of these capabilities.
As a rule, the phrase is used to describe systems that can configure, protect, optimize and heal themselves without a lot of input from the humanware that has until now been required to keep them up and running. (The term "autonomic," by the way, comes from the autonomic nervous system found in mammals and other higher order creatures - it refers to things like the heart beating and the function of the sweat glands - in other words, to those necessary body functions that we don't have to think about in order to perform.) At some levels the concept is new, but at least as far as storage is concerned, some of this has been with us for a while. Consider our RAID systems, for example. These have been with us since the late 1980s, and many of the more advanced examples have gotten pretty smart indeed. Their hot spares and redundant power supplies certainly represent an early example of autonomic computing. So does the smart cabinetry some vendors provide that can sense heat extremes, for example, and can then respond by cranking up their fans.
But while these are steps in the right direction, they are only first steps. Much more is needed before these devices are really autonomic, and certainly much more is needed before we can go home at night with the feeling that our systems are now pretty much tending to themselves.
Consider the case of storage resource management software. What should such management software have to do to be considered autonomic? What are the aspects of self-configuration, self-protection, self-optimization and self-healing that we should expect from them? This is something I plan to address over the next few issues. In the meantime, whatever that mix of needs that the autonomous system winds up addressing turns out to be, one ground rule ought to be clear: the system should be self-sustaining.
In this case the concept of self-sustaining system must refer not just to a subset of the overall storage environment, but rather to the entire system of storage assets that we rely on.