EMC, NetApp execs debate future of iSCSI

A discussion between an EMC and NetApp executive, via their blogs, has users and analysts debating: Will the "Year of iSCSI" ever really happen?

It's not necessarily news that EMC Corp. and Network Appliance Inc. (NetApp) disagree. These bitter rivals have been known to slug it out publicly over virtually every subject in IT. But as 2007 has gotten started, the subject du jour has been the hot topic of iSCSI, and its future. The debate between EMC and NetApp has sparked a lively discussion among users and analysts.

This latest round between the two heavyweights of the storage industry began with a blog post by Chuck Hollis, vice president of technology alliances for EMC.

In his post, Hollis cited IDC numbers showing that IP SANs have attained just 5% of the overall market despite years of heavy marketing pushes and predictions that "this is the 'Year of iSCSI.' "

More info on storage connectivity issues
Brocade buys Silverback for 10 GigE acceleration chips

EMC and NetApp duke it out in the midrange

ISCSI gains strength from servers

ISCSI gets a boost from InfiniBand
At least at the moment, Hollis wrote, "Despite the obvious economic advantages on the [iSCSI] hardware side, the people side of the equation not only negated any economic impact, but pretty much ensured that people who had started with Fibre Channel would largely stay with Fibre Channel for the time being. We could give the stuff away, it wouldn't matter."

In a follow-up interview with SearchStorage.com, Hollis expressed his frustration over what he views as slow adoption for iSCSI. "We as technologists often forget that just because it works better doesn't mean that users will want it," he said. Hollis said he has come to believe that Fibre Channel is too deeply entrenched in customer accounts at this point to ever be seriously threatened by iSCSI.

"There may be some organizational inertia," he said. "Customers are trained on Fibre Channel, they've been deploying it for years, and we in the IT industry who have been pushing iSCSI may have solved a problem they didn't think they had."

Meanwhile, in response to Hollis' post, NetApp's co-founder, executive vice president and resident corporate blogger Dave Hitz was -- perhaps predictably -- quick to take issue. "When I look at the iSCSI numbers for the past few years, what I see is an explosion of success," Hitz wrote, comparing the rising revenue for iSCSI products to NetApp's own quick rise to the top as a startup back in the 90s, a comparison that favors the growth of iSCSI. "ISCSI is now 10 times as big as NetApp was when we went public. I can tell you that it sure felt like the 'Year of NetApp' when we reached that milestone."

In the days since the discussion got under way, users and analysts have been taking their own sides. Most seem to agree with Hollis that the development of iSCSI into a widely used protocol has been slow, but hold starkly different opinions on what it means in the end.

In one corner: 'ISCSI will continue to languish'

Chris M. Evans, managing director for Brookend Ltd., a storage consultancy in the U.K., and a user/blogger who commented on Hollis' post, told SearchStorage he had seen the "organizational inertia" Hollis referred to firsthand. "When looking at Fibre Channel vs. IP [for one client] it was very difficult to get the network guys to give us a straight answer on what we wanted to do," he said. "In a lot of large organizations, networks and storage are separate teams, both with their own territory. ISCSI transcends that boundary somewhat, and I don't see either group giving up ownership of their part of the infrastructure."

There is no single vendor that defines iSCSI category, and until there is one, it will continue to languish.
Anil Gupta,
storage blogger and systems engineerQuantum Corp.
Another issue, according to Evans, is standardization. Relatively few Fibre Channel HBA, array and switch manufacturers, he pointed out, makes interoperability checking and certification more manageable. Meanwhile, having many more IP switch and NIC manufacturers makes it "impossible for any storage vendor to certify all iSCSI options in terms of external cards plus all the onboard NICs, so the stack is going to be inherently less stable," Evans said. "So iSCSI solves a problem in that it's cheap and simple to implement, but it will continue to lack mainstream uptake because of the support issues."

Another user, Anil Gupta, a storage blogger and storage systems engineer providing professional services and field support for Quantum Corp., expressed a similar opinion following the original debate on his own site. "The first four years of iSCSI revenue are divvied up between several iSCSI vendors with no single vendor dominating the iSCSI category," Gupta wrote. In contrast, he said, most network attached storage (NAS) revenue went to NetApp in the beginning, and NetApp was "equated with" NAS.

"Can you tell me which company is equated to iSCSI in its sixth year? Nobody." Gupta wrote. "There is no single vendor that defines iSCSI category, and until there is one, it will continue to languish."

"ISCSI has multiple battle lines to contend with" between NAS, DAS and Fibre Channel, according to Greg Schulz, founder and analyst with the StorageIO Group, who said he believes iSCSI will challenge Fibre Channel for greater market share, but not anytime soon. "Near term, while the hype will continue to be around iSCSI vs. Fibre Channel, the real battle will continue to evolve to be iSCSI vs. NAS, particularly in the midmarket, SMB, high-end SOHO and enterprise ROBO environments," Schulz wrote in an email to SearchStorage.

In the other corner: 'Time and again, Ethernet wins'

"People get impatient with iSCSI because its potential has been historically over stated, and it was somehow supposed to right the wrongs of the entire industry," said Brad O'Neill, senior analyst with the Taneja Group. But, he added, "Time and again, Ethernet wins -- the advantages of a common network are just too massive."

You don't instantly get the boulder to the top of the hill and claim this is the 'Year of the Boulder.'  You gain success with each passing day, week, month and year.
Tony Asaro,
analystEnterprise Strategy Group
"Think IBM OS/2 and Novell Netware 2 vs. Windows 3.1," said Tory Skyers, network administrator for Fox and Roach Realtors, the nation's fourth largest realtor and a subsidiary of Prudential Financial Inc. As recently as a decade ago, Skyers said, Novell and OS/2 were the standard operating systems, with huge corporate support that wouldn't dream of touching Windows, he said.

"Microsoft said, 'OK fine, you win -- we bow out of corporate space, we'll take over the user's desktop and own it,' " Skyers said. "Now iSCSI is saying, 'OK, fine, Fibre Channel, you win -- we will bow out of enterprise and take over the SMB green field.' " Eventually, according to Skyers, the Windows operating system took over at home, and because it was more familiar to corporate employees, edged its way into the enterprise as well. This, he predicted, is exactly what will happen with iSCSI.

"All those SMB greenfields grow up to become Fortune-whatever companies," Skyers said. Home storage use is exploding as well, he pointed out, and it's all being done over Ethernet. "Big companies will soon start hiring folks that use the iSCSI all their lives -- it is already happening."

"I just think we got caught up in big marketing statements like 'The Year of Blah' -- and it is silly," said Tony Asaro, analyst with the Enterprise Strategy Group (ESG), referring to Hollis' question of when the long-predicted 'Year of iSCSI' will come. "It's like pushing a boulder up a hill. You don't instantly get the boulder to the top of the hill and claim this is the 'Year of the Boulder.' You gain success with each passing day, week, month and year."

What do YOU think? Email Beth or comment in our Sound Off! section.

Read more on Networking hardware