From Evolution to Revolution: The Myth of the Fibre Channel SAN

Unfortunately, the vision of networked storage articulated by the ENSA white paper authors (and other pioneers at Sun Microsystems and elsewhere) was coopted and recontextualized by the marketing departments of early SAN vendors in the late 1990s and early 2000s. Like many 20th Century revolutionary movements, the marketing forces behind Fibre Channel were not content to wait for the supposedly inevitable and evolutionary force of some unseen hand to bring about the shift to networked storage. Rather, like so many revolutionary zealots, they sought to stir up the hype around their preferred technology, to make a case for early adoption, and to help the inevitable revolution along. In short order, Fibre Channel SANs came to be characterized as a Clayton Christensen “style disruptive technology that "revolutionized" storage.

Vendors seized on the myth of the data explosion to promote solutions that bore little or no resemblance to the ENSA storage utility (or the comparable SAN architectures promoted by Sun's Project StoreX). Fibre Channel was pressed into service by a powerful industry association as the "plumbing" of a SAN, despite the fact that the protocol was, at least at that time, incapable of producing a true network by any valid definition of the term . Advocates of the Fibre Channel SAN, an oxymoronic name for a switched fabric of point-to-point connections, quickly became the dominant voice in forums where networked storage was being discussed. Despite its inherent limitations, vendors offered the FC SAN as "the only game in town" for addressing burgeoning data.

One myth, that of a data explosion, thus fueled a second: the myth of the Fibre Channel SAN. Today, FC SANs are represented as a de facto standard in network storage ”one that separates storage from servers, and one that offers a real panacea for what ails enterprise storage with its nondisruptive scalability, universal accessibility, and improved management and security for enterprise data.

Analysts are contributing to the hype, offering that FC SANs are finding their way into the enterprise storage infrastructures of public and private organizations at a growing clip: Combined annual growth rates in the high 65 percent range are estimated for FC SANs through 2004. ^[6] By contrast, adoption rates for server-attached storage are tallied to be growing at "only" about 8 percent in the same period.

We could all learn a lesson about lying with statistics from the purveyors of these types of projections. Considering that the number of server-attached storage solutions deployed in the field is already enormous , and the number of deployed FC SANs remains quite small, analyst projections of comparative growth rates are deceptive. Metaphorically, if the Centers for Disease Control issued a statement that the number of persons infected with the common cold would increase by 8 percent this year, makers of cold remedies would delight at the prospect of millions of new customers. By contrast, if the CDC said that the number of persons infected by "Ebola Virus" would increase by 68 percent this year, the resulting number of new cases (about 10) would hardly create a murmur in the pharmaceutical industry and treatments for the disease would likely remain in the category of "orphan drugs."

So it is with FC SANs and server-attached storage platforms: According to some industry watchers, there were only about 11,000 FC SANs deployed as of 2001. ^[7] The preponderance had less than a terabyte of data, calling into question whether the FC SAN has achieved widespread acceptance as the de facto enterprise network storage topology at all. ^[8]

In the next chapter, we will further deconstruct the mythology around FC SANs and examine the evolution both of Fibre Channel standards and of IP-based storage networking protocols that portend to provide the plumbing for networked storage.