Microsoft’s cloud offering, called Azure, has been commercially released to the public only recently, though a community preview beta has been publicly available for longer. Thus, it is important to note that some elements of the Azure platform are likely to change in future commercial editions.

Like the Google App Engine, Azure is a PaaS offering.

Developers create applications using Microsoft development tools an Azure extension to the standard toolset allows such applications to be developed and deployedon Microsoft’s cloud, in much the same manner as developing and deploying using Google App Engine’s SDK. There are also similarities with aspects of Amazon’s IaaS offering, such as the use of virtualization, user control over the number of virtual servers allocated to an application and user control on elasticity. However, unlike the non-relational Google Datastore and Amazon SimpleDB, the recently released commercial edition of Azure provides relational storage services, albeit with certain limitations as we cover below. Azure also allows storage of arbitrary files and objects like Amazon S3 as well as aqueuing service similar to Amazon SQS.

Application code deployed on Azure executes in a number of virtual machines (called instances)with each virtual machine running the Windows Server operating system. While users do not have any control on when instances boot and how longthey stay up, they can specify the number of instances a particular applicationis likely to require.

From what was available in the community preview of Azure and its rapidevolution in the commercial edition (such as the addition of full relationaldatabase support, at least for small applications), it is likely to become eveneasier to migrate existing Microsoft applications to Azure in the future. Therefore it may well be that Microsoft Azure becomes a more popular PaaS platform

at least for large enterprises, as compared to Google’s App Engine.