Amazon's Elastic Computing Cloud (EC2) service, the virtual processing companion piece to its Simple Storage Service (S3), is officially coming out of beta status today, a transition heralded with the introduction of a reasonable Service Level Agreement, open-beta Windows support, and the promise of new features to come in the next year including automatic load balancing, scaling, and improved monitoring and management tools.

The announcement is widely seen as a pre-emptive move to steal the thunder from the expected upcoming introduction of Microsoft's own cloud-based processing service, but it's also a nicely timed move to gain inroads with enterprise IT during an especially vulnerable moment. With the economy in a shambles, cuts are on the horizon... and although the quality and value of cloud services is debatable, the cost savings are not. CIOs who might never otherwise have considered EC2 as an option may be spurred out of their safe, conventional approaches to bulk processing by massive budget cuts and be willing to take the chance on a migration to a cloud service. Amazon is gambling that they'll find the performance acceptable and stay with the service even as competitors come online and the economy recovers.

This emphasis can be seen in Amazon CTO Werner Vogel's companion post to the announcement, titled "Using the Cloud to build highly-efficient systems." Vogels reiterates, at a nearly kindergarten level for the CxOs who haven't yet got the message, how to do exactly that, a sort of primer on the way the world will be working in five or ten years.

The move to production status isn't much more than a publicity stunt, and I'm not saying that as a bad thing. The feature announcements are so much vaporware at this point, but whether they come out next year or not, they are certainly on the way, and there are third-party alternatives already extant to fill those holes. The SLA of 99.95% availability is just okay, but it's also of less concern than with S3, both because EC2 has historically had fewer availability issues and by it's nature it can be better distributed to avoid catastrophic failures at single chokepoints. The opening of the Windows hosting beta is more interesting; it probably had to be done now, before Microsoft's PDC introduction of their own version of a Windows cloud, but it answers quite a few of the questions about pricing and licensing that were raised when Amazon first announced the development of such service. Full time use of a basic Windows instance running 24/7 will cost you about three bucks a day... that works out to about nine months of use before you cross the point where it would have been cheaper just to buy the license yourself, excluding hardware and admin costs. Let's call it a year. But of course because you only pay for what you use, it's extremely unlikely that a year of processing time would actually be gone in a year... most activities in which you might use this service are burst-oriented, a lot of processing for a limited amount of time. Every company would have to do its own math, of course, but it's quite conceivable that using an Amazon EC2 instance of Windows for batch processing could be perpetually cheaper than buying and running a Windows box... and that's before you account for any scaling advantages you may accrue from the approach.

I think this is a good move from Amazon at a good time. As terrible as the economic downturn is, it may well accelerate certain inevitable trends in the technology industry, and for all the inevitable disruption that will cause, in the long term it's a positive thing.