Microsoft's Eating Caviar In The Cloud

One of my favorite terms in Microsoft-speak is "dogfooding", the act of using a product or service internally so that its developers can experience it the same way that customers will. When it comes to cloud computing, though, Microsoft seems to have lost its way. Their cloud efforts reflect an internal diet that's more like caviar than dogfood.Here's an example. This week, Microsoft is upgrading their MSDN blogging platform. For the next week, bloggers will not be able to make new posts unless they were queued before the upgrade started. Also, nobody will be able to make posts on new or existing blog entries. So here we have one whole week of downtime for Microsoft's communication with their users. Could any of Microsoft's own customers afford the luxury of a week-long offline upgrade for their own companies?

Another case is Windows Live Hotmail. An update is coming soon to Hotmail. Another blog entry talks about how Microsoft uses feedback from Hotmail users to decide which features to implement. This new release has a lot of "social" features, but still lacks the basic IMAP functionality that any decent mail system should have. Users hunger for standard IMAP connectivity, but Microsoft serves them the hacky Outlook Connector instead.

Both of these examples expose Microsoft's version-oriented and feature-oriented view of software. Rather than seeing the web as a series of services that are always online, Microsoft seems to have no problem with taking their blogs offline for a week, which is an eternity in Internet time. Rather than incrementally improving Hotmail with small tweaks over time, Microsoft focuses on big-bang releases and still manages to ignore the need for basic IMAP functionality.

Still not convinced? Here's another experiment you can try. Spend five minutes looking at the Amazon Web Services page. Now head over to the Microsoft Azure page with the same five-minute study period. Which offering do you think you understand better? Which one seems to reflect the needs of companies building web-based services? To me, Amazon wins on both counts. Microsoft's Azure diagram, no doubt exported from a lengthy PowerPoint deck, talks in vague generalities but never seems to really explain what is being offered.

Rumor is that Google will be following Amazon and announcing a similar set of services at their Google I/O conference this week. This is further bad news for the caviar consumers at Microsoft. Amazon and Google aren't coming up with some untested ivory-tower idea of what their customers might buy. Instead, they're selling versions of the same services that they use on their own web sites, ones that have been tested, proven, and refined over several years of real use. Now that's dogfooding.