What's hot, and what's not

One of our cover stories this month ("Analysts share 2002-03 storage perspectives") is based on a poll of some leading storage consultants and analysts. Among other things, we asked them to rank storage technologies (mostly emerging) according to how important they will be to end users in 2003 (see table on p. 18).

I weighed those results with the results of numerous polls of our readers (end users and integrators), threw in my own opinions, and here's my two cents:

Policy-based management software—I'd give this the highest-priority ranking for end users, along with related software management categories such as storage resource management (SRM) and applications-managed storage. The reason barely needs repeating: Hardware costs continue to decline, but the costs associated with managing those resources continue to skyrocket—whether you know it or not. Emerging storage management tools, which will eventually be integrated into cohesive suites, will help administrators contain costs—the number-one priority at most IT shops.

Standards-based management software (based on CIM and WBEM)—This will be a low priority for two reasons: Across-the-board vendor support probably won't happen until at least 2004 (despite early support from a handful of vendors), and even when it does become reality, standards-based management software will probably provide a relatively low level of management capabilities.

Disk-based backup—At least on paper, disk-to-disk backup offers an attractive alternative to traditional tape-based backup. But the argument isn't as simple as "disk is faster than (and almost as inexpensive as) tape." Nevertheless, disk-based backup will be a very high priority this year, although tape will continue to dominate for archival applications. (For more information about this, see the Special Reports in this issue.)

IP SANs—Godot might actually show up this year. The iSCSI spec will finally be completed within the next month or two; big players such as Cisco and Intel (and, soon, Microsoft with its free iSCSI drivers) are pushing the technology; even Fibre Channel stalwarts such as Brocade and McData have plans for multi-protocol switches; a slew of start-ups have interesting iSCSI products, at least for the low end; and it's hard to argue against taking advantage of your existing infrastructure.

Fibre Channel SANs—iSCSI-based IP SANs will take hold at the low end of the market, relegating Fibre Channel to the high end. Since that's where Fibre Channel has always been, encroachment by IP SANs will not necessarily bode poorly for Fibre Channel vendors, at least not within the next year.

Virtualization—Fortunately, I'm out of space.

Dave Simpson,

This article was originally published on February 01, 2003