In-Depth
Enterprise Storage: Storage Dollars and Sense
On February 13, the common stocks of major host bus adapter and controller manufacturers, as well as storage networking giants, dropped in value suddenly and precipitously. The reversal of fortune caught many financial analysts,
fund managers and stockbrokers by surprise. They scrambled to the phones seeking explanations from eminent industry analysts for the sudden decline in a technology sector that they had earlier represented as special, exceptional and unflappable. Two immediate responses were received:
• The stocks of most technology companies, including those of many storage vendors, were overvalued, observed some analysts smugly. It was only a matter of time before the market exacted its reality check.
• The general economic slowdown that affected other tech sectors was also affecting storage. Quoting the vendors, analysts now explained that companies were beginning to husband their resources, to keep equipment in production longer to delay purchases. This amounted to reduced demand for servers, software and storage.
Unsatisfied by these entirely cogent explanations, a few investment counselors sought opinions further down the pundit food chain. A few even tapped lowly storage industry columnists for their insights. The question was simple: Since the downturn in economic growth had not been accompanied by a slowdown in data growth, why weren’t storage vendor stocks continuing their upward climb on the NASDAQ ticker?
It was not all storage vendors whose stocks fell on February 13, but a select few with little in common, save for their involvement with Fibre Channel SANs. That the revenue forecasts for these companies had slowed came as little surprise.
In the Trenches
There is mounting evidence that many IT managers are reconsidering the deployment of SANs, based on experience with first-generation, Fibre Channel-based products. This is not a wholesale dismissal of the SAN value proposition: The benefits of a scalable, manageable, secure and intelligent storage infrastructure are widely understood. The SAN pioneers have successfully framed the issues, raised awareness and garnered mindshare, regarding the importance of the networked storage infrastructure.
The problem is that today’s SAN has not solved all of the issues that networked storage was intended to solve. Relying on the same three dollars and sense categories originally used to make the case for the SAN – cost reduction, risk reduction and new business enablement – IT managers are now reexamining the fit for current-generation products to their specific problems and strategic objectives.
Of particular concern to present day, budget-wary IT managers is storage cost of ownership. To reduce the cost of storage, the IT manager needs to reduce the cost of storage management, which is estimated by various analysts to total between $17 and $35 per MB, per year. SAN vendors claim that SANs provide a centralized storage infrastructure that avails itself of better management by fewer personnel. Deploying a first generation SAN in order to centralize storage, however, is not the only approach – and not necessarily the most dependable, reliable or cost-effective one – for building a centralized storage infrastructure or for tackling storage management costs.
Many IT managers are rightfully concerned about the lack of "in the wire" or in-band management capabilities in the current generation of Fibre Channel. Until management fabric services are added to the protocol (this will happen in 2002, according to the Fibre Channel Industry Association’s technology roadmap), managing storage in the Fibre Channel SAN requires either (1) the use of some proprietary in-band method that doesn’t necessarily work in a heterogeneous SAN or (2) slinging a second network – an Ethernet network, in most cases – to handle device management. This dual-network topology inhibits the ready scaling of SANs, by the way, calling into question the cost-efficiency of the infrastructure, overall.
An alternative to SAN-based consolidation is to use trustworthy "Big Iron" storage arrays that provide effective methods for sharing data among multiple hosts. Sticker shock sometimes accompanies the $450,000 to $600,000 starting price for an enterprise class array. However, a Fibre Channel SAN – at $1,400 per switch port – is often no less expensive.
Another option is to use Network Attached Storage (NAS), a technology for connecting storage devices directly to the production LAN. This strategy is flexible, enabling either physical or logical storage
consolidation. It also facilitates centralized management via standard network management protocols already pedigreed for enterprise-class use. NAS solutions range from the very inexpensive to prices rivaling the Big Iron arrays, but they may have definite cost of ownership advantages over first-generation SANs.
Jon William Toigo is an independent consultant and author of The Holy Grail of Data Storage Management. He can be reached at [email protected].
About the Author
Jon William Toigo is chairman of The Data Management Institute, the CEO of data management consulting and research firm Toigo Partners International, as well as a contributing editor to Enterprise Systems and its Storage Strategies columnist. Mr. Toigo is the author of 14 books, including Disaster Recovery Planning, 3rd Edition, and The Holy Grail of Network Storage Management, both from Prentice Hall.