This issue includes the second installment in our four-part special series on business continuity and disaster recovery. BC/DR has always been a top IT priority at large organizations. And two years ago this month, it became a priority for all organizations. Last month's electrical blackout again focused attention on business continuity.
Steve Kenniston, an analyst with the Enterprise Storage Group, takes a look at a variety of technologies that can help you align your data-protection requirements with your business continuity objectives, including disk-to-disk backup, virtual tape, snapshots, continuous data capture, and good old replication. (See "Data protection still comes first," p. 22.)
Phil Goodwin, an analyst with the META Group consulting firm, likens developing a disaster-recovery portfolio to developing a personal investment portfolio in "Are you ready for a 'recovery disaster?' ", p. 26. He provides sound DR investment tips and outlines a tiered strategy for data protection and recovery. Unfortunately, he notes, budget cuts have left many IT managers with a Clint Eastwood DR strategy: "Do you feel lucky today?" Well, do you?
Closely related to BC/DR, security issues are raising their ugly heads at many IT sites today. That's nothing new. But in the past, security issues were largely the domain of network and server administrators. However, as more and more storage becomes networked, storage administrators are getting into the security act. Tom Petrocelli, a consultant with Technology Alignment, tackles the subject in "Guidelines for ensuring storage network security," p. 36.
Tom makes some fundamental points. For one, remember that most security attacks come from within, not outside, an organization. And instead of waiting for the slick security tools and boxes that vendors are promising in the next year or two, focus on what you can do today.
If you're a regular reader of InfoStor, you're aware of the potential promises of IP storage, and iSCSI specifically. You're also aware that the iSCSI industry has not delivered on any of those promises—until very recently.
Two articles in this issue take different tacks on the iSCSI market. Senior editor Lisa Coleman's article ("Distributors report low demand for iSCSI," p. 8) shows that, despite all the hoopla over the last three years, resellers and end users aren't exactly knocking down the doors to get to the nirvana of iSCSI SANs. That conclusion is based on interviews with distributors that have a vested interest in selling new storage solutions.
For a more positive view of IP storage developments this year—including iSCSI, FCIP, and iFCP—see our monthly column from the Storage Networking Industry Association, by David Dale, p. 46.
Storage vendors and disk drive manufacturers spend a lot of time debating interface technologies. So do storage integrators.
End users don't really care: They just want fast, capacious disks at affordable prices. In our Opinion column this month, The Storage Consulting Group's Richard Lee weighs in on the Serial ATA vs. Serial-Attached SCSI debate.
In his view, Serial ATA—despite its demonic acronym, SATA—is the disruptive technology that will shake things up. For end users, the advent of SATA is clearly good news because it provides "good-enough" storage at affordable prices.
It's show time
There's a lot of debate about which storage conferences are best, but next month you can check out one of the biggest—Storage Networking World, Oct. 27 to 30, in Orlando. For more information, go to www.storage networkingworld.com.
To the Editor
(Editor's note: The following was sent in response to the editorial in the August 2003 issue of InfoStor, p. 6.)
Do the big storage network players want adoption of 10Gbps or 4Gbps Fibre Channel? If they go to 4Gbps they don't keep up with Ethernet, which is moving to 10Gbps. If Ethernet moves quickly to 10Gbps and Fibre Channel slows 10Gbps adoption by allowing 4Gbps along the way, IP-based protocols for storage networking will gain the legitimacy required to challenge Fibre Channel as the dominant storage interconnect protocol sooner rather than later. With the amount of time and energy that has been put into Fibre Channel, I'm sure equipment makers, OEMs, and even end users who have adopted Fibre Channel do not want the protocol to lose its competitive edge.
Whether an end user chooses to stop at 4Gbps or 10Gbps, there is going to be a problem. 10Gbps Fibre Channel has no migration path. As currently designed it will not be possible to attach a cable from a 1Gbps or 2Gbps port to a 10Gbps port. That means rip-and-replace or co-exist. However, it doesn't make a difference. While 4Gbps products will work with 1Gbps and 2Gbps products, they will not work with 10Gbps products. So in the eyes of Fibre Channel proponents all 4Gbps does is slow down adoption of 10Gbps, leave the same migration issues, and further minimize Fibre Channel as a protocol relative to IP-based protocols.
The move from 1Gbps to 10Gbps Ethernet provides the same connectivity problems. However, there are few, if any, production SANs using 1Gbps Ethernet. So no integration problems. To Fibre Channel proponents, that represents equal footing. For greenfield opportunities the decision then comes down to reliability and price. Fibre Channel loses.
Steve Berg, CFA
Punk, Ziegel & Co.
In the July 2003 cover story, "Focus shifts from backup to recovery," we incorrectly classified StorageTek's EchoView as a next-generation snapshot technology. EchoView does not use any snapshot, nor does it create snapshots. In fact, EchoView behaves more like a continuous capture technology, so it should have been listed as a "real-time recovery" product.