I attended a briefing on Microsoft Azure, its cloud platform. We were given details on Azure pricing and availability, along with brief presentations from organizations intending or hoping to make use of it. The press release is here.

Microsoft is promising commercial availability in Q4 2009, possibly at PDC 2009, scheduled for Nov 17-19. This will be English only, and will cover 21 countries including the US, UK, Ireland, Canada, India, Japan, Australia and New Zealand. In March 2010 availability will be extended to over 30 countries, though Azure itself will still be English language only.

Pricing is as follows:

• Azure: $0.12 per hour per VM. Storage $0.15 per GB/Month plus $0.01 per 10K transactions.
• SQL Azure: Web edition with 1GB database for $9.99 per month, or Business Edition with 10GB database for $99.99 per month.
• .NET Services: $0.15 per 100K message operations.
• Data transfer: $0.10 per GB inbound and $0.15 per GB outbound.

It is hard to compare directly; but my initial reaction is that Microsoft is somewhat more expensive than Amazon Web Services. SQL Server looks especially pricey; and I wonder if Microsoft is conscious of all its partners selling SQL Server as part of a hosting offering and not wanting to undercut them. Another legacy problem.

Microsoft has also announced SLAs for the new service, promising 99.9% uptime for Azure services with a 10% credit in the event of failure (that’s my summary of a slightly more complex guarantee). This is useful as a statement of intent, but not particularly valuable from a business perspective. The cost of downtime can be huge, and a 10% service credit does not begin to compensate.

I found the case studies interesting, even though they were generally expressions of intent rather than absolutely firm plans. Perhaps the most compelling was EasyJet, who told us that airport systems are generally antiquated and inefficient, and plans to bypass them as far as possible using an Azure-based alternative. If I can find time I’ll write this up in more detail.

Judging by these case studies, Azure is appealing to Microsoft-platform customers who enjoy being able to take their existing or new ASP.NET applications and migrate them to Azure with relatively few changes. Without the abandonment of SQL Data Services in favour of full SQL Server this would not have been possible.

Is Microsoft serious about Azure? I’m beginning to think that it is, though the marketing has been muted, and I am sure the company is aware of the tensions inherent in selling server licenses for on-premise use on the one hand, and services which make them unnecessary on the other. I doubt whether Microsoft would be in this space at all, if it did not fear the likes of Google App Engine eating away at its business.