Much has been written about iSCSI's promise for small and medium-sized organizations, due to its familiar IP-based technology and comparative low cost,
In fact, "the majority of the market success to date, mostly through the efforts of NetApp, has ended up in the enterprise space, because that's where they [NetApp] already had a presence," says Arun Taneja, consulting analyst and founder of the Taneja Group.
Over the long term, however, Taneja says he suspects enterprise customers will only use iSCSI selectively, where needed. By contrast, "in midsized companies, where they haven't yet made a commitment to Fibre Channel, they will gravitate toward starting with iSCSI for their SANs."
James Opfer, research vice president, Gartner Research, also singles out NetApp as the company that has done the most to move iSCSI implementations into high gear so far. And, like Taneja, he sees plenty of opportunity for iSCSI. But Opfer also sees different market motivations. For one thing, Opfer says is not convinced that iSCSI is always less expensive than FC when lined up feature for feature. However, "on the server side, the real play is with the Windows market, where the iSCSI driver is free, NICs are inexpensive, Ethernet switches are probably already available and it is inexpensive to connect additional servers," he says.
With that market segment in mind, Opfer foresees an evolution similar to that experienced by FC in the late 1990s -- namely starting where the number of servers isn't large, but where there is a good opportunity to separate the storage from the server. In the Windows space, "the first opportunity is Exchange," Opfer says. In cases where there is more than one server, administrators can use iSCSI to separate storage from the server, and use Ethernet for backup.
What's more, Opfer says an analysis of FC adoption worldwide shows it is clustered around the most expensive servers. "In places with less expensive servers, it is virtually invisible," he says. And while he does not expect FC to stand still (more aggressive product pricing, for instance, is already starting to materialize), the bottom line is that there is a huge opportunity for iSCSI to serve market segments that FC has never before touched, which means in terms of total deployments, "iSCSI will overtake FC, and pretty quickly," Opfer says.
But there will be no commanding leads or knock-out blows in this battle. Each technology will have adherents for a long time to come. "The real interesting thing to watch on iSCSI is 10 gigabit Ethernet and how that plays out," he adds. Opfer says the problem is at the 10 gigabit level "the NICs aren't free anymore," and the economic advantages may fade. And, since Gartner is not expecting that to happen until 2007, "that gives Fibre Channel plenty of time to get even more entrenched," Opfer says.
For more information:
About the author: Alan Earls is a freelance writer in Franklin, Mass.
This was first published in March 2005