Professionalism, Pragmatism, and Consumerism

As has been suggested throughout this book, there is no "secret sauce" technology that can deliver an intelligent storage infrastructure today. The intelligence in any storage infrastructure is a function of effective design and pragmatic technology choices.

The basic building blocks of storage, especially disk drives , are becoming more and more commoditized. It is intellectual property ”primarily in the form of software and firmware ”that is increasingly used by vendors to differentiate products from one another.

Intellectual property is the true wellspring from which new ideas flow from the drawing board to the research and development laboratory and ultimately manifest themselves in products themselves . Vendors have a legitimate argument to make when they want to be compensated for their product IP. It is not the argument of this book that all storage is commodity storage. Storage products can be differentiated by their added-value features and functions, and the market has its own way of deciding the worth of innovation.

Problems begin to arise when products are represented as so different from one another that any comparisons between them are viewed as fundamentally invalid. The claim of many large-scale array vendors is basically that "When God created my product, he broke the mold."

This is not true, and the vendors know it. However, with substantial marketing budgets to shape consumer perceptions, with storage managers themselves effectively isolated and atomized, and with analysts and media feeding from the troth of vendor marketing and advertising budgets , who is going to question the vendor line?

The situation becomes worse when you realize that vendors are the primary participants in standards development organizations. On the one hand, it makes a great deal of sense for the best and brightest technicians to help develop baseline standards for technology. However, the flip side is that standards development is too often held hostage to the marketing objectives of the vendors themselves.

In the most severe case, this takes the form of a declaration by the chairman of the IP Storage Working Group at IETF that his company already holds a general patent on all IP-based storage networking and that the work of the group will eventually need to pass through his company for licensing before it is implemented in anyone else's gear. ^[9] Another case of excessive vendor control is seen when a large networking company leverages its market presence to obtain adoption of its proprietary SAN technology ”just so it can claim to be standards-based. ^[10] In still other cases, vendors water down standards so there is enough "wiggle room" that even their most proprietary and noninteroperable products can claim to be standards compliant. ^[11]

One could respond that the market will shake loose all pretenders in time, as it did in the case of TCP/IP and Ethernet protocol-based products over a 20-year period. Consumers will ultimately vote with their wallets.

However, this assumes that consumers have 20 years to wait. Moreover, it takes for granted that consumers are informed about products, their capabilities, and limitations, and about alternatives. That is assuming a lot.

Surveying the literature of the storage industry, one finds precious little in the way of product performance test data. Vendors have in some cases set up roadblocks to prevent the dissemination of unsanctioned test results. One major array vendor goes so far as to invalidate its warranty if a customer discusses publicly the performance received from its product.

As for due diligence testing by consumers themselves, this is less and less frequently done for reasons of time, lack of facilities, and lack of knowledge. As mentioned at the outset of this book, test labs appear to be among the first targets of corporate cost-cutting in lean economic times. Without them, it is difficult to do your own evaluation.

It should come as no surprise that an increasing number of consumers outsource their product decision making altogether to resellers and solutions integrators who often serve suppliers with greater enthusiasm than they do their customers. During the course of the development of this book, I had to turn down a lucrative speaking opportunity from a top value-added reseller. The company wanted me to tell their reseller/integrator clients that Fibre Channel SANs were a tremendous opportunity to cash in and that they would be doing their customers ”and their own bank accounts ”a favor by recommending Fibre Channel for all storage requirements. That the customer didn't really need a SAN in every case was beside the point: the objective was to fire up attendees (resellers) about Fibre Channel SANs so they would help move a large inventory of product before "legacy" Fibre Channel was replaced by IP storage.

While there are doubtless some trustworthy resellers and integrators, outsourcing storage decision making is generally unwise. As a matter of best practice, resellers can be tasked to test solutions and to offer other advice, but never should they be allowed to assume the role of technology decision-maker for a company. This practice makes about as much sense as the practice of buying products on the basis of vendor reputation and tenure alone.

In the current economic client, it is common to see big storage consumers pass on the purchase of technology from any vendor whose name is not three letters long (e.g., IBM, HDS, EMC, etc.). Their rationale is typically one of fear: They are afraid that the smaller vendor will disappear and they will be left with an unsupported product.

However, given the short payback interval large companies expect from their storage acquisitions (that is, the abbreviated period of time in which cost-savings realized from the product are expected to offset acquisition costs), this philosophy really makes no sense. Typically, in larger companies, a technology acquisition is expected to return its acquisition expense within about 13 months. Moreover, a replacement for a product is already being sought within a year of product implementation. So, the total exposure that a company faces to an unsuccessful product acquisition is about one year or less. By contrast, many technologies offered by both long-tenured vendors and relative newcomers return their investment within days, weeks, or months of implementation.

Established vendors have been quick to capitalize on consumer conservatism and to raise the spectre of doubt about start-ups with innovative ideas. At about the time of this writing, a brand-name vendor took the opportunity as a keynote speaker at an industry tradeshow to announce the results of the testing that his internal lab had performed of the disk-to-disk backup products of several start-up companies. The speaker used the results, which were arrived at through undisclosed methodology, to dismiss the products of burgeoning competitors outright . Doubtless, a negative impression of the new products was left with all of the consumers in attendance, coming from such a reputable source.

These shenanigans are not confined to established players. In 2001, a newcomer in the NAS space pronounced that its product was a hundred times faster than industry leading products from established players. The vendor presented internal test data and compared it with published speeds and feeds data taken from competitors' websites to make his case. Later, it was revealed that the tests measured only how quickly data could be retrieved from the memory cache of the newcomer's storage appliance, while competitive statistics reflected the amount of time required to retrieve data from the back-end disk subsystems of the competitors' arrays ”in short, an unfair comparison.

What's worse: Employees of the same newcomer vendor later misrepresented themselves on a web-based bulletin board frequented by investors interested in a competitor vendor's stock. The perpetrators claimed to be customers of the competitor's storage products and went to great pains to criticize the architecture, performance, and service delivered around the vendor's products. Asked whether management was aware of the employee behavior, the response from a company spokesperson was that management had been aware of "one or two instances" and had since instructed staff "not to engage in such actions in the future." In any other industry, such behavior would have been grounds for dismissal.

While vendors engage in such campaigns against each other, the consumer is victimized by poor information. In the worst case, they are held in outright disdain by vendors. During one exchange over standards version numbering conventions for the iSCSI protocol, a participant on the IP Storage Working Group, representing a prominent storage technology company, stated unabashedly that it didn't matter which numbering scheme was used to identify the version of the protocol. He noted that his customers didn't know how to use a protocol analyzer anyway, so they couldn't look at the version number and determine if their equipment supported the final standard or some prestandard variant. Veracity of the statement aside, it underscored the concern of this author that vendors do not answer to the consumer anymore.