the whole premise of microsoft cloud technologies is windows, windows virtualization, and .net. from what i understand things like lamp, ruby on rails, and so forth reign supreme in this industry. is ballmer on crack?
Don't underestimate the gobloads of legacy apps based on microsoft tech that are still running in the enterprise. There are a lot of developers out there with a massive investment in MSFT tools & knowledge. Giving enterprises a chance to run those legacy apps in the cloud is a viable strategy. Sure, it's nothing for the 'cool kids', but why should MSFT care if money can be made? But yeah, 'largest cloud computing provider' they will never be.
The issue is primarily licensing, .NET is not ASP. It's actually a very good platform for development. The problem however is cost. Ballmer is very much not on crack because the same guys who buy Visual Studio will be willing to pay MS's price for cloud computing.
Yes, RoR and LAMP reign supreme on the cloud right now because most of the cloud toolsets and tool chains are geared towards RoR and LAMP. The cloud is very inaccessible to .NET developers of which there are a lot and they are employed by companies with large budgets who can afford such things.
When Ballmer gets the "Deploy to Windows Azure" button working in Visual Studio he will be on a gold mine. Especially since a lot of the people who will be clicking that button write horribly inefficient code. Instead of knowing their code is crap because the application is god awfully slow it will instead just ring in lots of CPU and IO cycles which result in big bills. The result will be "thats just how much an applications costs" instead of "write better code"
good points. i think cloud is a broad set of technologies in which operating system, database, web server, apis, etc play a part. msft could make some inroads with these as services considering their past record for shrink wrap software. but for ballmer to say that they will dominate the cloud, that's a bit of a stretch imho.
i feel that msft is trying to play the incompatibility card again where if they make things easier to develop in their set of cloud languages and technologies, it'll pull more developers and (unfortunately) bring us back into the days of internet explorer hell once more. for me, if i'm trying to build up a highly scalable web application, if it takes less time to write in msft jargon then perhaps lamp/ror becomes less attractive, but i haven't seen the day yet.