I'm having a hard time interpreting the results of an end-user survey on storage area networks (SANs), conducted by Enterprise Management Associates. EMA surveyed 158 readers, and the results are both perplexing and revealing.
Consider the responses to the following question regarding SAN vendors: "From which vendors have you purchased or do you plan to purchase SAN hardware?"
Although Cisco will create a tsunami in the SAN pond next year, the company currently is not a SAN vendor. So why did Cisco show up on the list? There are a few possible explanations, but I think the most likely is that end users don't define "SAN" the same way that storage vendors do.
Most vendors and industry analysts use the term "SAN" to describe a separate network dedicated to storage. SANs are-for the time being-based on Fibre Channel technology. But the inclusion of Cisco in this list implies that end users do not use the same definitions because Cisco does not make Fibre Channel hardware. And it will be at least another six months before Cisco has a Fibre Channel blade (developed in conjunction with Brocade) for its Catalyst switches. So, definitions could be one explanation: What we have here is a failure to communicate.
Another possible explanation for the inclusion of Cisco in the list of the top six SAN vendors is that end users think the future of SANs rests not in the hands of Fibre Channel, but in Internet Protocol (IP) networks such as Gigabit Ethernet. Now, in the Fibre Channel community, that's hot potato heresy, but in the end-user community it may be a wish come true. If the IT community moves toward IP-based storage (most likely via the iSCSI standard under development) then, of course, Cisco will be one of the SAN leaders.
The Fibre Channel versus iSCSI controversy will pick up steam as iSCSI moves from paper napkins to shipping products, but for now I'd like to stick to the Network World survey.
Another thing that popped out of the survey results is Network Appliance's inclusion in the list. NetApp is the leader in network-attached storage (NAS), but the company is not a SAN vendor. Again, we may have a definition problem. Most vendors and industry analysts agree that the main difference between the two architectures is that NAS operates at the file level and SAN operates at the block level. A further differentiation-at least for now-is that NAS attaches to Ethernet and SAN is based on a Fibre Channel network.
But based on the survey, end users apparently don't differentiate between the two: Why else include Network Appliance as a primary SAN provider?
My point in bringing this up is not to haggle over definitions. More importantly, the results of surveys like this call into question other survey results, particularly the ones that gauge end-user adoption of SANs. For example, if 29% of the respondents say they have already installed a SAN (which was the case in this survey), how many of those are actually NAS implementations and/or Cisco environments?
To be sure, a number of (Fibre Channel) vendors have pointed out to me the shortcomings of the survey: a limited response base, the network (as opposed to storage) orientation of Network World readers, etc. Those objections aside, IT managers who are trying to gauge the rate of SAN adoption among their colleagues should note that not everyone's on the same definition page.
Dave Simpson, Editor-in-Chief
The questionable quadrant
In another highly debatable report, the Gartner Group IT consulting firm recently produced one of its notorious quadrant rankings (see chart). Compaq published the report at www.gartner.com/webletter/compaqstorage/october/index.html.
Gartner ranked Compaq and EMC as the SAN leaders after weighing a wide set of factors that boil down to "ability to execute" and "completeness of vision." The results of this "study" are not as surprising as the Network World survey, but they're just as debatable.