Dispelling the Myth of Cloud Standards
There's a lot of talk about getting cloud standards, but one look at the ecosystem and you'll realize talk is all it is.
With the rapid development of cloud computing technologies, it is perhaps no surprise that discussions have already arisen regarding standards for cloud computing. But "standards" in this case may equate to "the technology my company is using" and as such is no more than mere hype.
Trying to examine the shape of cloud standards is like trying to trace the shape of a cloud at any given moment: amorphous and constantly on the move. Several vendors have stepped up to offer their own cloud "standards," and there's even little agreement on what organization will even manage cloud standards, as groups like the Distributed Management Task Force (DMTF), Cloud Security Alliance, Open Grid Forum, and the National Institute of Standards and Technology have all formed committees and working groups to work out the details of cloud standards. And that's the short list.
Any casual observer of the cloud community will immediately pick up on the presence of Amazon's Elastic Cloud Computing (EC2) platform as a very dominant player in cloud space. With EC2 such a popular consumed platform, they could argue, isn't it becoming the de facto standard?
There's some truth to that, because in terms of sheer number of cloud users, EC2 is likely only second to Google App Engine as a cloud services provider. The fact that this is not widely known is testament to Amazon's market positioning: Despite being number two in users, Amazon has definitely generated more mojo than Google in this space.
But being number two, or even number one, may not be enough. The standards argument is more nuanced than a numbers game. At their core, standards deal with the API level, where application developers typically tie their applications to the platform. The EC2 APIs are good, but their features are very much dictated by Amazon Web Services, not as part of a true standard.
The EC2 APIs are not open in any sense of the word. While the Amazon Software License lets developers routinely tie into the APIs, you can only use such applications on--you guessed it--Amazon's cloud. And vendors that are offering EC2 comparability, such as Eucalyptus and Cloud.com, are always going to be playing catch-up to the EC2 feature set until Amazon sells true licensing to these vendors.
What seems apparent in the cloud space now is that EC2 clearly got the perceptual jump on all the other cloud providers, who are now left to scramble around to get caught up and somehow slow EC2 down enough that they won't win the industry standard by default.
It's easy to liken this situation to the old Betamax vs. VHS argument: Where one vendor's "standard" (Sony's Betamax) was eventually brought down by the concerted efforts of other vendors to deliver another standard (VHS) on cheaper, commoditized machines. It did not matter that Betamax was arguably better than VHS--the market voted with their wallets and Sony failed to respond in time.
Is a similar scenario about to play out for Amazon? Nothing's impossible, but right now it seems unlikely that Amazon has anything to worry about. The cloud ecosystem, perhaps because of its immaturity, is in great disarray and the smaller players seem unwilling to band together as they focus on getting their own piece of the cloud pie. One might look to Google for stepping in to lead a competitive standard, but the company seems unwilling to start such a movement. Maybe it's content making more money than $Deity on everything else it's doing.
Amazon is playing this perfectly right now: it has not done more than pay lip service to any of the cloud standards organizations, and it's been sparse lip service at that. By staying above the fray and letting the market vote with its feet and wallets, Amazon seems to be communicating it wants no part of any standards discussion, because it would be a waste of time.
Breaking the impasse
Two recent events might shake this plan up a bit. In late August, Red Hat submitted its open source Deltacloud platform to the DMTF as a possible interoperability standard.
Deltacloud is not new; it was announced at the 2009 Red Hat Summit in Chicago. Deltacloud creates an abstraction layer that cloud application developers can tie into instead of natively touching each cloud provider's API. According to the Deltacloud site, developers "do not directly link a Deltacloud library into [their] program to use it. Instead, a client speaks the Deltacloud API over HTTP to a server which implements the REST interface."
The Deltacloud API, in turn, uses drivers that plug into the each of the supported cloud providers' APIs. This technology would be a huge advantage to cloud developers, since it would go a long way toward preventing lock-in. Any app that hits the Deltacloud layer could be easily migrated to any cloud provider.
The other event that might affect Amazon's strategy is the advent of the OpenStack cloud platform, which uses open APIs in its infrastructure. If OpenStack catches on, Amazon might find itself watching customers migrate away from EC2 to get features the EC2 APIs might not have. Rackspace, OpenStack's primary sponsor, certainly has the infrastructure to handle a big cloud consumer base.
Looking at all of these power plays for customers, it seems clear that any talk of standards is premature. Cloud feature sets haven't been fully developed yet, and the volatility of the marketplace seems a poor environment for any real discussion of standards.
Something to keep in mind the next time you hear a vendor tout the "next cloud standard."
Brian Proffitt is a Linux and Open Source expert who writes for a number of publications. Formerly the Community Manager for Linux.com and the Linux Foundation, he is the author of 18 Linux and Open Source books, including his most recent work, Introducing Fedora: Desktop Linux. Follow him @TheTechScribe on Twitter.