On Demand Storage Stressed at Comdex Page 3


Want the latest storage insights?

Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure

Autonomic Computing

Another speaker from IBM laid out another wave of technology innovation that is destined to impact the storage landscape — autonomic computing. IBM has essentially folded this term into its on demand strategy. The term autonomic basically means self-governing or independent. It is based upon the concept of the autonomic nervous system — the body governs heart beat, blood flow, glands, etc. without the person having to pay any attention to them.

"The amount it costs to manage the infrastructure is now more than the cost of the infrastructure itself," contends Rick Telfer, director of autonomic computing at IBM. "We need autonomic systems to bring about simplicity."

Essentially, autonomic computing heralds intelligent open systems that:

a) Manage complexity

b) Know themselves

c) Continuously tune themselves

d) Adapt to unpredictable conditions

e) Provide a safe environment by being self-configuring, self-healing (fixing problems as well as determination of causes), and self-optimizing (tuning systems to the workload at hand)

Telfer stresses that such technology would not eliminate the jobs of database administrators and storage administrators. "By automating the environment, you remove all the grunt work and manual entry," he explains. "That helps these administrators do their real jobs and perform their proper functions."

Other fears about on demand were addressed by Charlie Boyle, director of N1 architecture at Sun. "N1 does not mean that Hal takes over the organization tomorrow," he said.

He preferred to address the concept in terms of today's realities — systems with utilization rates that average anywhere from two to 25 percent. The goal of N1 is to be able to manage the entire storage landscape or data center from one system — i.e. virtualization that links storage to business processes, services, or specific servers depending on needs.

"The reality, though, is that most companies are not ready for such an advanced architecture," says Boyle. "Existing IT infrastructures are typically too complex and not matured enough to be able to transition easily."

Even so, he touts several organizations that have implemented the technology in a rudimentary form. One expects to realize three-year savings of $10.2 million with an online transactions system, while another has experienced a reduction in time spent on service provisioning from one week to a day.

"While the gains are real, there are many unknown processes that need to be addressed and sorted out inside any organization before you will see real business value," concludes Boyle.

Lori Wigle, director of the enterprise platform group at Intel, concurs. "On demand is no grand solution," she says. "If the basics are out, you cannot achieve a virtual stage. Yet many companies today don't know how many servers they have or where they are situated."

She sees these various on demand initiatives as a means of eliminating the complexity of storage and I/O. Blades, she believes, are a great platform for autonomic capabilities.

"Self-healing and self-optimization are a good four years away," predicts Wigle. "I believe it will take four years for true virtualization to become a reality."

Feature courtesy of EnterpriseIT Planet.

» See All Articles by Drew Robb

Submit a Comment


People are discussing this article with 0 comment(s)