Thoughts on Azure

At their annual PDC conference on Monday, Microsoft announced the CTP of their new cloud services platform, called Azure , code named “Red Dog”.

When I first heard about the offering, I immediately thought if they play their cards right, this could start a massive push of cloud services into “regular” corporate enterprises.

Disclosure: I’ve known about this for a while because Dell Data Center Solutions is providing hardware and services for the build out of Azure. In fact, the code name Red Dog has led DCS to have all sorts of K9 derived code names for products. We’ve been hounded with so many bad puns; I almost needed to flea the office the other day. (sorry)

Some initial reactions might be “yawn” – another cloud platform, isn’t everything a cloud platform now? What’s different with Microsoft?

First:

Visual Studio (and its ancestors) is an excellent IDE that is easy to use, especially for non-developers like me. I took a few programming classes in college (back when Windows 1.0 shipped), but not a whole lot of QBASIC has been used in the past few years. But I have been able to work my way through a variety of things on my own.

I’ve always been amazed at what a “real” developer can do. And that’s where Azure comes in. Someone who really knows what they are doing will be able to integrate Azure services into the applications they are writing without having to go out of their way. Everything will be right in the Visual Studio IDE. This means that the massive number of companies developing on the Visual Studio platform will instantly have access to these services in a relatively easy manner.

I’m also hopeful that a set of Azure plug-ins for eclipse will be developed, which will spur development in the open source community.

Second:

Most cloud vendors don’t have large enterprise sales teams. Early cloud customers have been self-selected: they had a problem and did the research to find the right vendor to solve that problem. Microsoft has a sales force on the ground in thousands of companies. And you can bet that they have marketing funds available to help those companies launch projects that integrate with Azure. It’s pretty hard to put a human face on the sales force of most cloud vendors. While that might be some of the appeal of the cloud, large companies still prefer to purchase things from a human.

Azure also brings up the interesting topic of cloud portability. One of the big selling points for the cloud is that you can pick up and move to another provider. But we’ve always known that wasn’t really true. Each vendor has their own APIs and just like traditional application environments, once you buy into a framework, your costs to move from that framework are huge. This concept is perfectly framed by a discussion on the Google cloud computing group. One poster said:

“Unless their [Microsoft Azure] storage API is S3 compatible, I don't see us moving 1,000,000+ objectSleep from AWS [Amazon Web Services]… am I the only one who considers parity with the AWS API a must to switch?”

This statement not only discusses portability, but cloud standards. Are the Amazon S3 APIs the de-facto cloud storage APIs?

Now – before I’m accused of drinking the Kool-Aid, here’s the skeptic in me:

· Perception is reality. If there are trust issues with the security of Microsoft products, especially if the service has a security failure early on, then there will be big problems. It can be hard enough securing company data when it’s inside the firewall.

  • Microsoft 1.0 products have always had their…challenges. But they usually get it right after the first couple of service packs.
  • Reliability. Google and Amazon have had relatively few outages, but when they occur, it causes quite a stir. A lot of people will be watching to see how Microsoft fairs in this area.

Is it going to be wait and see or are you going to run with the Red Dogs in this new found land? For me, it’s time to get some chow. I know, these puns are terrierfying…

About the Author: Eric Kuzmack