Legacy system upgrades keep users up at night

14.12.2006
We tend to think of interoperability in product terms: Product A from vendor X works or doesn't work with product B from vendor Y. However, a survey conducted by the Storage Networking Industry Association's End User Council (SNIA's EUC) show that interoperability has just as significant a vendor support component.

The SNIA EUC surveyed 400+ storage/IT administrators from a broad spectrum of small/medium/large enterprises, across an equally broad spectrum of industries. The findings have just been released in a report entitled "Storage Interoperability: So what's the problem?"

Those of you in the IT trenches will probably not find it surprising that upgrades to legacy systems are most feared in terms of their nail-biting potential. Rapid nail-biting applies equally to legacy software and firmware upgrades, with "forced upgrades [by a vendor] to maintain support" following closely behind.

You may also nod your head at this one: The three most often cited causes for disruption are misconfiguration, human error, and non-disk related hardware failure. In other words screw-ups-by both the users and vendor support personnel-are just as disruptive as things actually breaking. So what fosters screw-ups? Some EUC members think the underlying complexity of today's storage infrastructure fosters mistakes, miscalculations, and misconfigurations on both sides of the aisle. In other words, if this stuff wasn't so darn complex, it would easier for all of us to support.

This survey also shows that the hardest storage networking component to get right is storage management software. Perhaps no surprises there either. More than half of the survey respondents report spending between 100 and 1,000 or more hours implementing storage management software, with roughly 15 percent in the 1,000-hours-or-more category. Daunting.

Finally, when asked to rank order a list of ten potential factors that would influence storage/vendor product selection for a product to be integrated with an existing SAN, "support reliability" was numero uno. Sadly, "external analyst evaluations" were also on the list. We finished dead last.