Friday, October 31, 2008

Azure in theory is great, in practice, not sure.

For those who don't follow this kind of thing, Microsoft announced and deployed the first version of their "cloud computing" offering called Azure this week.

Azure is really different than EC2, Amazon's "cloud computing" offering. Here's the difference:

  • EC2 is essentially like a big VMWare system. You take an image of an entire Linux server, or entire Windows server, and start it up. It runs just like a computer in a rack somewhere would run.

  • Azure abstracts that further. You write a piece of C# code and it runs somewhere, out there, and concerns at the OS level that you'd have with EC2 aren't necessary.

There's more power in the EC2 system if you know what you're doing and have the time. You can customize your infrastructure to whatever you need. I could use EC2 immediately with third party, licensed software like Mental Ray, for example.

Academically, I prefer the pure theory of Azure. It's a lot like Google App Engine... write some code and it just scales using the existing scalability of the company's infrastructure. And Azure has some signifcant advantages over Google App Engine ... C#, the best IDE out there, a more robust API. Plus, Microsoft has a lot more invested in the success of Azure than Google does the App Engine. So I don't think it's just something that was thrown out there on a whim kind of like GAE was (sorry guys, given the lack of progress that's just how it seems).

That said, Microsoft has a larger challenge than Google. The knock on GAE is that it requires developing for a platform that's not easy to get off of. Microsoft has the same knock, and they're additionally Microsoft. The reluctance of people to jump on board something proprietary to Microsoft may be high. On the other hand, if it can be shown that deploying on Azure is much faster than other solutions, it might not be such a difficult thing to sell. After all, Salesforce has a proprietary development environment that people seem ok with.

No comments: