SRM Gets The Gas
In Vendor Land, it's a short hop from capacity planning to storage resource management (SRM). A couple product guys from IBM volunteered to explain why this makes good business sense (even if it blows your budget).
March 4, 2008
In Vendor Land, it's a short hop from capacity planning to storage resource management (SRM). A couple product guys from IBM volunteered to explain why this makes good business sense (even if it blows your budget).As I mentioned yesterday, the expansive nature of any capacity planning conversation became abundantly clear while reporting this story. But John Foley, market manager for IBM's TotalStorage Productivity Center (TPC) software, and his colleague, Jason Davison, TPC product manager, tried to explain why it was useful to blow out the capacity planning discussion to a broader data management perspective.
Regardless of how broad you believe the conversation ought to be, you can't know the scale of your problem or challenge without a good bit of discovery, according to Davison. "If we can't manage or discover all the servers, switches, and storage systems, then the product doesn't have value for the user since they'll have to do it manually," he said. So collecting that info -- down to the SAN devices and hosts -- is the first step in "discovering" the entire data center.
From there, users need to be able to generate reports in any format, with any frequency or metric(s) they require. TPC has more than 300 report types, which one hopes covers most the major bases and then some.
Then users can begin to discern how much total capacity they use, how much is allocated and utilized, at what percentage it's growing, and how soon current capacity will be maxed out.
"The critical word there is 'trending'," said Foley. "If it's growing 10% to 15% a month, how quickly will I run out of storage? We show you historical and projected data and when you'll run out of storage at this rate."
"The age of that data is also critical," Foley said. And here's where we go from talking about capacity planning to SRM.
On the server side, TPC goes into the files themselves and reports on duplicate files, the age of the data, and the last time it was accessed. "Anything that hasn't been accessed in a year might be a good candidate for archiving," Davison explains.
"We also give performance charts with throughput for switches, the I/O, and cache hits on storage systems to see it from a trends perspective, not just what it's doing right now," Davison added. Spreadsheets and written records don't offer that.
So in addition to capacity planning, SRM also handles topology discovery, and resource visualization from the application to the LUNs on the storage box, as well as capacity and performance management; change and configuration management; and best practices guidance.
With the advent of virtualization and tiered storage (think Serial ATA versus Fibre Channel), these kinds of functions -- beyond capacity planning -- become more critical. "What sets [SRM] apart are emerging and differentiating features like data classification, like information life cycle management; availability and root cause analysis; audit control; task automation; and performance analytics," Foley said.
Foley and Davison also emphasized that customers don't need a full-blown SRM solution for it to be useful -- TPC can do a basic health check on an entire storage system, or it can be used to measure if service level agreements are being met.
"Without SRM, you've just bought a Ferrari without a gas gauge or speedometer," Foley added. But just how broadly enterprises buy into this sports car analogy remains to be seen.
Read more about:
2008About the Author
You May Also Like