In October, with great fanfare, the Open Data Center Alliance unfurled its banners. The ODCA is a consortium of approximately 50 large IT consumers, including large manufacturing, hosting and telecomm providers, with the avowed intent of developing standards for interoperable cloud computing. In addition to the roster of users, the announcement highlighted Intel with an ambiguous role as a technology advisor to the group. The ODCA believes that it will achieve some weight in the industry due to its estimated $50 billion per year of cumulative IT purchasing power, and the trade press was full of praises for influential users driving technology as opposed to allowing rapacious vendors such as HP and IBM to drive users down proprietary paths that lead to vendor lock-in.
Now that we’ve had a month or more to allow the purple prose to settle a bit, let’s look at the underlying claims, potential impact of the ODCA and the shifting roles of vendors and consumers of technology. And let’s not forget about the role of Intel.
First, let me state unambiguously that one of the core intentions of the ODCA, the desire to develop common use case models that will in turn drive vendors to develop products that comply with the models based on the economic clout of the ODCA members (and hopefully there will be a correlation between ODCA member requirements and those of a wider set of consumers), is a good idea. Vendors spend a lot of time talking to users and trying to understand their requirements, and having the ODCA as a proxy for the requirements of a lot of very influential customers will be a benefit to all concerned.
But it is a huge jump from acting as an honest broker to encourage things that almost everyone connected with cloud computing eventually wants – interoperability standards, common interfaces, common management models, etc. – and an organization with the de facto power to control purchases and dictate vendor product plans. In this role I think the ODCA will fail.
Consider some of the underlying background and industry realities:
- Standards in a rapidly evolving technology are difficult – The history of IT screams this fact at us. Group attempts at standardizing rapidly evolving technologies have usually been dismal failures, and true standardization has been driven by the Darwinian give and take of the marketplace. Sometimes de facto standards remain vendor controlled de facto standards and get widely deployed anyway. The reality of today’s technology marketplace is that technology is evolving faster than any standardization effort could evolve. Additionally, standards always lag proprietary practice because complex technology always works best in proprietary integrated stacks. Only after it has had some time to bake in the market can it be standardized. Standards groups gave us X400 and Token Ring, while the market gave us universal TCP/IP and Ethernet.
- Users are not necessarily the ones to define standards – End users of infrastructure technology as a group are excellent at defining what their problems are, and solving them with available technologies, but are not in the business of defining future products, architectures, or developing standards. The skill sets are different, and most IT groups, even in large companies, do not have people sitting around with nothing to do but try to become product architects. Additionally, consortiums of diverse corporations are not monolithic, and will make pragmatic buying decisions, not ones circumscribed by the group standards. If a “nonconforming” product is the best one to get their job done, that is the one that will be purchased. On the other hand, I&O groups are excellent at defining their problems, and the development of use cases and requirements is a strength of the teams, and this is the area where the ODCA can shine.
- The $50 billion of market clout is grossly exaggerated – This figure appears to be the total IT spend of these companies, and possibly with a little compressed air added. The actual portion of this total that is directly applicable to the cloud-related IT spend of the companies is hard to pin down, but is much, much smaller. Let’s say maybe 5% at a very generous maximum, leaving us with a couple of billion dollars distributed over 50 companies. A lot of money, but way short of $50 billion. And any leverage is years in the future, since the group hasn’t actually delivered anything yet.
- Evil vendors – Let’s face it, evil vendors are a fact of life. They are in it to make money. For themselves, at the expense of their competition, with tactics that often push to the very limits of what is allowed under law. That’s capitalism. And because of this, it has been the vendors who have pushed technology as far and fast as it has gone, and who have created things that most users never even dreamed of – ubiquitous compatible desktops, virtual machines, cloud computing models, the commercialization of the web, and so on ad infinitum. Without evil vendors we might still be writing on paper ledgers with quill pens. Admittedly, vendors sometimes need a little discipline from the courts, but all in all it's hard to think of a better system for promoting progress, and somewhat disingenuous to complain about proprietary lock-in with products that have been purchased with an eyes open evaluation of their price versus benefits. If the benefits did not outweigh the price, including the lock-in factor, they wouldn’t get purchased for long.
- Intel as the sole technology advisor – Intel’s role in the ODCA is not in any way altruistic, nor is it required to be. Intel will almost certainly at a minimum use their relationship with the ODCA to gain early access to emerging requirements so they can factor them into their product decisions and where possible gently steer things in directions more compatible with what Intel wants to do. This is not in any sense evil, but rather behavior well within the cultural norms of the technology industry, and Intel has a lot of smart people who can probably contribute to the group’s efforts. But we would be more comfortable with the ODCA’s overall credibility and ability to deliver vendor-neutral products if AMD was a member, along with an assortment of other vendors with cloud portfolios.
Netting it out, the ODCA has a strong potential to develop use cases that can be of value to the vendor community, which will in turn encourage vendors to build products that deliver the desired solutions to customers. In communicating these requirements, and in providing a ready channel to market for vendors who produce conformant products, the ODCA can provide a service. But I think the odds are against it becoming the kind of powerhouse gating organization with immense market clout that its organizers projected in its opening days.