Back in 2003 much was written about storage resource management, with products from Softek, Trellisoft, CA and others. Customers could get much better storage uitilisation by using SRM products, and defer buying expensive storage. But these ideas turned out to be impractical and then three technology developments stopped SRM in its tracks.
SRM found itself flailing and left behind as the world took on board ideas of information lifecycle management (ILM) and compliance. Latterly we have all realised, or re-realised, that storage virtualisation is a necessary underpinning of any storage management software. This all meant that SRM vendors were faced with massive re-writes of their software or rip-and-replacement of software components. No wonder things have gone quiet.
Softek CEO steven Murphy says: "I think SRM has stalled. People are relying on their primary disk vendor for this tool."
Softek bought itself out of Fujitsu in April, 2004. It then gradually understood the prospect that revenues from its suite of SRM products, with its visualisation functions, were not going to deliver what it wanted. What was Murphy to do? Plans to have an IPO were put on hold. A comprehensive review of its software assets was undertaken and the company realised that its data migration technology was a terrific asset.
It has a strong presence in the mainframe world with data migration and data replication products. Softek is good at getting data bits off one disk array and onto another. Not a commonly required product you might think; data migrations being few and far between. Actually you'd be dead wrong.
A moving target
Softek ran studies of data migration. It found that among 256 end users in the Europe, Middle East and Africa area, 57 percent migrated data on a weekly, monthly or quarterly basis. Migrations were complex; a quarter of the users took more than four weeks to plan a migration. They were risky; a staggering 82 percent reported problems when migrating data; 39 percent of them reported technical incompatibility issues.
Softek had found a pain point. It was worse in North America; there was a 20 percent greater occurrence of data migration problems.
Softek asked the Enterprise Storage Group to run an independent study. Like Softek, ESG found that data migrations were a fact of life in the IT world. More impressive findings were revealed: 83 percent of them exceed the originally planned staff time; 61 percent exceed the planned downtime; and over half go over budget. This was a real pain point. Murphy reckoned Softek had proven products and technology that could accomplish non-disruptive data migration, migration with no downtime at all.
A new direction
This is where Murphy is focussing the company: "We're very strong in the data mover area. Our real go-to-market strategy is where data is migrated. We know exactly what we wear now and we fit it like a glove. We have probably the highest performing asynchronous data transport. This is the core, not the visualisation stuff."
"Ninety percent of migrations are done at the weekend. Most are over time, over budget and exceed the planned downtime. The fact that we can do it non-disruptively means they can do it in the week."
Most migrations are manual, intensively manual. Murphy asserts that it just has to be automated.
A new product?
Softek sells a data mover, TDMF, and replication in the mainframe world. TDMF is used daily for load-balancing in the mainframe world. Softek's replication technology is now available in the open systems arena, meaning Windows and Unix.
The new direction for Softek and Murphy's comments lead Techworld to speculate that TDMF functionality will become available on Windows and Unix systems, such as AIX, HP-UX, and Solaris, perhaps Linux. What it will probably offer is migration between volumes in a storage array and migration between arrays. It might even appear before the end of 2005.
Softek is a see, move, and recover company now, Murphy says. You can use its products to see where your data is located, move it between volumes and arrays, replicate it asynchronously across distances for disaster recovery purposes, and, of course, recover the data.
SRM future directions
The SRM products of 2003 aimed high and fell short. SRM products of the future will probably come from developing storage utility vendors. They will be based on a foundation layer of virtualisation technology with a set of storage services - protection (of various sorts), provisioning, information life cycle management, compliance archiving - layered on top and a management interface acting as a single pane of glass to all the facilities.
In our view, don't expect any real products to emerge until 2007 or 8.