The Storage Industry: Fear of Standards and Commoditization


The Storage Industry: Fear of Standards and Commoditization

In point of fact, data storage remains one of the few areas of information technology characterized by a paucity of standards ”whether open or de facto . In this, data storage stands in stark contrast to other technology sectors, where well-defined standards have either:

  1. Been articulated by open standards organizations like American National Standards Institute (ANSI), Internet Engineering Task Force (IETF) and adopted across literally all vendor product lines, or

  2. Imposed by a single vendor with a commanding market share leadership position (e.g., de facto standards).

An example of an open standards-based technology is the network interface card (NIC). In the NIC market, all manufacturers must adhere to a set of formal open standards if they want to sell their products. As a result of open standards, end-users can pick and choose products from the more than 100 products and still achieve their basic goal: connection to an Ethernet network operating the TCP/IP protocol suite, the dominant network protocol in the world today.

Because of open standards for Ethernet and TCP/IP, any NIC card will do the basic job of providing network attachment. Customers can, with confidence, buy a NIC from any vendor, install it in their desktop computer or server, and be sure that their system is ready to be connected to a network.

Of course, to differentiate their products from those of their competitors , some vendors "add value" to their NICs by adding features such as "Wake on LAN" or special TCP offload engine (TOE) chips, but the basic standards-based functionality is available whether the consumer buys the enhanced NIC or its less expensive "bare bones" cousin.

In the final analysis, consumers are the beneficiaries of open standards because such standards level the playing field between vendors, enable the "apples-to-apples" comparisons of different products, and provide customers with assurances of baseline compatibility that, in turn , frees them to consider options for purchase on the basis of feature, function, and price. The presence of open standards in the network space is owed to the work of a generation of network technology start-ups who sought to unseat a previous generation of proprietary network technology vendors, and also of a unified consumer voice that voted with their checkbooks for open standards-based products.

Standards breed commoditization, of course. And commoditization of technology exerts a downward pressure on product prices. That is why a gigabit Ethernet switch that cost $3,000 per port in late 2000 dropped precipitously to about $450 per port in late 2001 and today hovers at around $300. As more standards-compliant switches came to market that provided essentially the same business value, most customers preferred products that were less expensive to comparable products that were more expensive. The result of this purchasing pattern has been price erosion , another feature of open standards-based technology and one that nearly always works to the benefit of the consumer.

Another example of a technology segment "blessed" (from the consumer's standpoint, at least) by standards is the PC market. As of this writing, about 80 percent of personal computer operating systems (OS) bear the logo of Microsoft. The Windows OS is a de facto standard in desktop computing. Of course, Windows was never formally ratified as an open standard by an official standards body, but, by sheer presence in the market, the Microsoft OS dominates all contenders.

The result of a de facto standard is similar to that of an open standard in at least one respect: It fosters compatibility. The dominance of Microsoft, in turn, drives most independent software vendors (ISVs) to develop application software to run on the platform. For example, this tax season , an end- user can go to a local computer software store and select from among several personal tax preparation software packages from several ISVs, any of which will run on his or her Windows-based PC.

Whether or not you prefer the Microsoft platform is beside the point. The fact that nearly every desktop application software package will operate under the Windows operating system means that the application software products compete on a level playing field imposed by a de facto standard. Vendors must add features and functions to their products, or offer licenses under an attractive pricing scheme, if they want to compete. Again, the beneficiary of competition in a standards-based realm is the consumer.

Storage lacks both open and de facto standards. And the industry resists virtually any effort to impose meaningful standards (whether by a standards-making body or by one of its own members ) that might compromise proprietary advantage for the vendor. Insights offered by many storage industry insiders underscore this position.

For example, a representative of one of the industry's largest disk array vendors tells the tale of how his management's biggest fear is that there will be a repeat, in the storage array business, of what happened in the disk drive business in the mid-1990s. In the late 1980s and early 1990s, disk drives featured proprietary technologies ranging from vendor-specific interfaces and low-level drive formats to specialized actuator arm and read/write head designs. This diversity ensured that Brand X products worked very differently from Brand Y and that consumers would be locked in to a particular vendor's technology. As a consequence, drive prices remained very high and margins were lucrative for disk drive manufacturers.

With the arrival of the Small Computer Systems Interface (SCSI) standard, however, this picture changed. SCSI provided a standards-based mechanism for communicating with storage devices and had the effect of standardizing drive electronics and interfaces. SCSI was not the only standard for disk drive interfaces ”other standards emerged at about the same time, including IDE/ATA. However, the overall impact of standards was to reduce disk drive products to the level of commodity or stock items.

Prices eroded dramatically as low-cost, standards-compliant drives flooded the market from mass producers . According to the source, the greatest fear at his company ”a large integrated disk array manufacturer ”is that it could happen again. Originally, the taxonomy offered by researchers at the University of California at Berkeley to describe various Redundant Arrays of Independent Disks (so-called RAID levels) was suspected of introducing a commoditizing effect in storage arrays. However, as documented in The Holy Grail of Data Storage Management , vendors treated the UC-Berkeley definitions less as a standard taxonomy than as a useful point of reference for describing their varied implementation of disk redundancy schemes. Some even coined RAID levels to describe their products that weren't in the UC Berkeley taxonomy at all. However, this isn't the end of the story.

The same fear of standards-based commoditization has echoed and re- echoed in the public statements of leading storage industry spokespersons over the past couple of years . At a storage networking conference in spring 2001, the then-CEO for a market share leading company in storage arrays noted that the reason his company was transitioning from a hardware-focused to software-focused model was to leverage the only differentiator that remained to distinguish his products from those of his competitors: namely, intellectual property captured in the form of software on his company's arrays. Without software, said the CEO, there was no difference between his products and those of his nearest competitor: "We both just sell a box of Seagate hard drives." That statement underscored the fear of lost differentiation in the market as storage hardware became less proprietary and more commoditized.

As standards drive commoditization, vendors know that they must find new ways to realize proprietary technology-based margins from otherwise commoditized hardware technology. Sometimes, this takes the form of resistance to standards efforts altogether.

According to a spokesperson for a Fibre Channel "SAN" [4] switch-maker in 1999, "A real storage area network ”one based on truly open standards and capable of supporting heterogeneous storage arrays (i.e., arrays from different product manufacturers) ”will probably never appear in the market. If a real SAN did appear, Joe's JBODs (an abbreviation for Just a Bunch of Disks) would perform exactly the same as high-end, 'name brand,' proprietary arrays, which cost ten times as much. The ' name brand' vendor's customers would quickly realize that they had been paying way too much for storage." [5]

"Under those circumstances," the fellow continued , "do you really think that the leading vendor would allow his arrays to be included in an open SAN?"

The question crystallizes much of what has Balkanized early SAN standards development efforts, as well as bringing into sharp focus the concerns that vendors have about open standards and their commoditizing effect on products.

Another Industry Concern: The Rise of a Dominant Player

Arguably, the only thing that strikes more fear into the hearts of storage technology companies than open standards-based product commoditization is the fear that one of their peers will achieve such market dominance that it will be able to impose its own technology as a de facto standard on the entire industry ”   la Microsoft in the desktop world.

Evidence of this concern can be inferred from the brief history of the Storage Networking Industry Association (SNIA). SNIA is a quasi-standards group comprised of vendors of storage technology products and services ”the brainchild of several well-intentioned (and possibly idealistic) storage industry analysts and insiders who perceived a need, in the absence of standards for storage networks, to provide a forum for "co-opetition" (a blend of cooperation and competition) among storage vendors.

Among other things, SNIA was set up to foster informal agreements between vendors on their implementations of networked storage technology so that the entire storage networking initiative would not fall prey to petty vendor infighting. The interoperability difficulties in the Fibre Channel SANs of the late 1990s were proof positive to the founders that some sort of SNIA-like body was needed.

The fact is, however, that SNIA never received the full backing of the industry until a major storage vendor, EMC Corporation, announced its own competing alliance of companies, the Fibre Alliance, to advance the vendor's own view of how storage networks should work. Rather than allow a single vendor achieve the status of a broker of de facto standards, EMC's competitors turned their attention to SNIA and joined up with the fledgling organization in droves. [6]

This phenomenon repeated itself in 2002 when a number of vendors, troubled by the unilateral move of one of their peers (EMC, again) to advance its own model for universal storage management, decided to counter the initiative by supporting a common management approach under development at SNIA: Common Information Model/Web-based Enterprise Management (CIM/WBEM). Until that time, vendors had paid only lip service to CIM, making no real provisions for CIM-based management in their own platforms or products. However, the fear of an emerging de facto standard, in the form of EMC's AutoIS initiative and WideSky "middleware," compelled vendors to "take the pledge" at the Storage Networking World conference in Palm Desert, California, in Spring 2002 and to promise CIM-manageable platforms by year's end! [7]

Whether or not the industry keeps its pledge to embrace a CIM-based management standard remains to be seen, of course. Cynics suggest that once the specter of single vendor dominance in universal storage management recedes, CIM developers at SNIA will see considerable backsliding among the newly converted. No one, after all, really wants a universal storage management standard.

In the past, those who suggested that consumers might benefit from an open cross-platform standards-based storage management methodology were met by cold stares from storage manufacturers ”reminiscent of Anthony Hopkin's character, Hannibal Lecter, in the film version of the Thomas Harris novel , Silence of the Lambs . At the risk of belaboring the metaphor, in one scene from the popular film, the hero, a female FBI agent named Clarice Starling, passes a questionnaire to Lecter that is intended to gather information about the serial murderer's psychology for the FBI's behavioral science database. Hopkins, as Lecter, levels a cold, unwavering stare at Starling and asks in a calm and calculated tone whether she believes she can really "dissect" him with such a blunt instrument (i.e., the survey).

This is exactly the response that one could have expected from a storage manufacturer prior to the spring 2002 CIM announcement: It was heresy for anyone to suggest that there was enough common ground among all of the 17-odd thousand storage platform products in the market that would enable their management by a common management methodology such as CIM. [8] The very idea was as heretical to storage vendors as it once was to Compaq and other PC and server manufacturers when the Desktop Management Task Force (DMTF) first tried to apply the CIM strategy to PC management in the early 1990s.



The Holy Grail of Network Storage Management
The Holy Grail of Network Storage Management
ISBN: 0130284165
EAN: 2147483647
Year: 2003
Pages: 96

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net