Microsoft and Amazon have been the two primary players in the Cloud Computing market for a while – the market for externally hosted public cloud, anyway.
They haven't been directly competitive, though, despite what they and nearly everyone who writes about them says (including me).
They do complain about each other and try to outdo each other, but each is aimed at a very different group of potential customers.
Amazon is an Infrastructure as a Service company, meaning you hire them when you want a real, live data center available that allows you to do all kinds of customization and optimization and performance tuning and network configuration and resource allocation and virtual-infrastructure design and maintenance. There is no hand-holding allowed.
Some of your eyes just glazed over; the rest of you are thinking "more detail, please."
For the first group there's Microsoft Azure – a Platform as a Service offering which, if you prefer it that way, allows you to build or upload applications built in .Net onto what looks a lot like one or more Windows Servers that you can manage remotely.
Almost since they launched, though, each has been sprinting toward the other by adding features to make its service more like the opposite one.
That's good if each has exactly half the features a single, largely homogeneous group of potential customers wants to have.
It's bad if they offer services that are different enough that the "weakness" of each is actually a strength for the customers or functions for whom it's designed.
I keep thinking they're already as close as they're going to get, but it's not true.
Amazon started out, essentially, renting an open-plan data-center platform for applications, then gradually shrank things down until customers could buy power in chunks almost as small as those offered by Azure.
Azure loosened up its single-image approach to give customers more control over the virtual infrastructure, where the resources were assigned, and even parallel processing for ersatz high-performance computing in the cloud. Oh. And Cobol, although no one is sure if that's really a benefit or not.
Now Amazon is announceing a product called CloudFormation that lets customers create a template (think Platform) that makes it easier to treat all the resources involved as if they're part of a single system.
Which is great, unless you're one of those people from earlier who wanted More Detail Please.
Microsoft is trying to recreate the strategy that won it the desktop by attracting as many application vendors as possible to its platform -- extending an offer for free time on the cloud for developers, making development simpler and letting them start with very small virtual machines rather than larger chunks of resources.
Surveys from Quest and Forrester late in 2010 estimated as many as 40 percent of companies have no plans to add cloud computing to their IT menus.
That means 60 percent do, which is an enormous market. And it's getting more crowded.
NetSuite and Informatica are challenging both of them.
HP is challenging...well, it's never clear who HP is challenging, but it's doing a lot of quality cloud-ish stuff.
Everything is rushing to the cloud, which means there's going to be a lot of room there for big competitors, and a lot of potential to stake out space and get really good at what you're trying to offer.
Amazon and Microsoft keep trying to go head-to-head, though, even though that's not where their heads or their customers should be.
More flexibility in your cloud provider is always a good thing, especially when the market is young and a lot of the functionality is undeveloped. Better to have providers round out their offerings than get hidebound too quickly.
These two should forget about each other a little bit, though, and just focus on doing what they do better, not try to knock out the other guy.