Microsoft announces “cloud in a box” appliance

Microsoft is taking an appliance approach to cloud computing with the announcement of a “cloud in a box” version of its Azure platform. At its Worldwide Partner Conference today, Microsoft described it as “the first turnkey cloud services platform for deployment in customer and service provider datacenters”.

Early adopters of a limited production release include Dell, eBay, Fujitsu and HP. Since the Windows Azure platform appliance runs on Microsoft-specified hardware, three of these will also be suppliers. This leaves eBay as the only real customer, and it already uses Azure.

However, by far the biggest user is Microsoft itself. The company has already set up some of the world’s largest data centres, with one serving Europe based in Dublin. It has also tried building giant data centres out of systems pre-installed in shipping containers, so it does literally have “cloud in a box” technology, albeit rather large boxes by most standards.

Microsoft is building data centres to run online services such as Hotmail and SkyDrive, its own business applications such as Exchange and CRM, and any applications that customers will pay it to run. Third parties can also offer Azure clouds for rent, and customers can build their own clouds if they want. For this, Microsoft is offering a Private Cloud Deployment Kit.

Bob Muglia, president of Microsoft’s Server and Tools Business, claimed that “Microsoft is the first and only company that offers customers and partners a full range of cloud capabilities and the flexibility to deploy these services where and how they wish — whether that is with Microsoft, a service provider, in a customer datacenter or a combination of all three.”

By a strange coincidence, this approach could also maximise Microsoft’s revenues by getting people to pay for Windows on their PCs, on their in-house servers and, again, in the cloud. Rather than payments for cloud services reducing the profitability of Microsoft’s Windows and Server divisions, Microsoft hopes it will increase them.

In theory, of course, companies can only get the maximum cost savings by getting rid of their own data centres and moving everything to the cloud, but it’s not clear how many large companies will be willing to take such a big risk. Some may also be concerned about the countries in which their data are stored and various regulatory issues. The mix-and-match approach enables enterprises to control the risk while testing cloud services to see how much money they can save, and how far they are willing to give up customisation and control.

The fact that Azure is based on Windows Server and Microsoft SQL Server should also simplify the transition for companies that already use those products in their data centres.

Today, Microsoft also announced a second beta version and pricing for its InTune cloud-based service while extending availability to regions of the USA, Canada, Mexico, Puerto Rico, France, Germany, Ireland, Spain, Italy and the UK.

InTune provides automated support for business PCs, including anti-malware services, for $11 per PC per month. The subscription includes an upgrade to the Enterprise version of Windows 7.

It’s not clear whether Microsoft hopes this will develop into a significant business (its consumer-oriented OneCare service failed) or whether it will mainly be a product for partners who already offer remote support services.

The InTune announcement on the Windows Blog quotes partners such as JCF Managed Solutions, Microload and Enfo Zipper to the effect that InTune will enable them to support more customers and expand their businesses.

It’s now a decade since Microsoft first announced cloud services (including a hosted version of Microsoft Office) and net-wide single sign-on and gateway services (Hailstorm). At the time, the sales pitch was for ASPs (Application Service Providers) who would operate much like ISPs but offer SaaS (Software as a Service). It even had instant-on thin (or thinnish) client “cloud computers” called Web Companions, which ran Microsoft Windows CE on ARM chips, rather than ChromeOS. Perhaps that effort was a bit ahead of its time. Well, it tanked. Microsoft should do better now.