Can storage hardware companies, used as they are to competing with each other on all fronts, be trusted to develop and implement standards for storage management? Some might suggest the very idea is like turkeys voting for Christmas, but not Scott Kennedy, the vice president of business development at Fujitsu Softek.
Softek's business is storage resource management (SRM) software that is independent of hardware. Kennedy argues that the benefits to be gained from emerging SRM standards such as SNIA's SMI-S and CIM-SAN are too great to ignore, even for hardware companies, and says that the standards process is close to bearing fruit too.
"Some of the scepticism is because the hardware companies came up with brilliant software because hardware revenues were falling and they needed the higher margins of software," he says. "But hardware-specific CDs are one of the most inefficient ways to manage, and they heighten the management cost because employees have to learn hardware-specific tools.
"The question is who do you trust? But [the hardware vendors] are not just dipping a toe in the water - there's way too many people working on it. We're not at the inflection point yet but we're really close."
To show the sort of problem that storage throws up, Kennedy tells of meeting a Korean customer of his, just as the latter got a report showing that only 11% of his storage infrastructure was unused. Initially thinking that a mere 11% wastage was very good, Kennedy was surprised to hear that the system was only a few months old and really should not already be 89% full.
"He couldn't figure out why the system was crashing every three weeks," Kennedy says, adding that the SRM tools helped track it down to the root partition filling up. It turned out that the company's Oracle installation had been set up wrongly and was leaving its dump files there, causing the system crashes.
SMI-S and CIM-SAN are already able to fix some SRM issues, he adds. "We saw encouraging things at Storage Networking World in Phoenix, for example basic steps such as discovery, LUN creation, zoning, map and mask. Advanced functions are going to be step two."
The use of standards means that although it takes the same amount of work to build a management tool, that same tool can manage any standard storage and does not have to be reworked for new hardware or manufacturers.
For example, Softek's SANview storage discovery software is built on family modules for different hardware, so the company created a family module for SMI-S as part of a mass demonstration in Phoenix. "We turned off the other family modules to see what we could detect, and we were able to detect 14 or 15 devices," Kennedy says.
"We also put a SMI-S provider on our box to make any storage standards-compliant even if it's not, so then the customer can use an SMI-S management tool.
"But then there's the 30-step process of putting a LUN on a server - SMI-S is designed to push that down to the device level. The slow-down point is questions such as: is this a raw disk? Am I expanding a volume? What are the nuances between operating systems that allow them to use a modified device without a reboot?"
Why isolation won't work
All of these are major challenges to those developing the standards, but Kennedy says that the storage hardware manufacturers are finally realising that they cannot carry on in isolation much longer.
"Disruptive technologies such as Serial-ATA will be a real awakening to the large storage box vendors. It's running at one-eighth the cost of large Fibre Channel boxes, so a Petabyte of disk becomes affordable," he says.
"Space, performance, availability, recoverability - if an array has all those, the market will adopt it. So then customers have ten times the storage and they still aren't managing it well! Storage management is spending a little bit to save a lot."
With that in mind, he sees nothing but long term growth for the SRM business, as long as product planners can keep their nerve and fight off accountants obsessed with cutting short-term costs.
"BMC exiting the market with Patrol is the silliest thing I've ever heard," he says. "That product was solving business needs! The idea of closing that group down makes me think the CFO [chief financial officer] had got control of the road-map."