When Microsoft first rolled out its Windows Azure cloud computing architecture, it stepped a bit out of persona, going so far as to support work on a Java SDK for non-C# developers looking to place apps on Microsoft’s new cloud. Still, most updates aimed at easier Azure cloud application deployment are of the .NET variety – not Java or J EE.
Exceptions to the ”.NET-mainly” trend for Azure have emerged. Earlier this year, interoperability specialist JNBridge released JNBridge Pro 6.0 with support for cross platform cloud implementations that span the .NET and Java languages. This week, Java in-memory data grid pioneer GigaSpaces announced tools that take complex Java application and integration software as-is and places it on the Windows Azure cloud platform.
Known as Cloudify for Azure, the software prepares applications by providing a Groovy-based domain-specific language for bundling deployment scripts, as well as basic out-of-the-box patterns for launching Java-on-the-cloud elements that can include the Apache server, Cassandra distributed database, the Spring framework, the XAP in memory data grid and others.
The developer can work in a familiar Java environment, which distinguishes Cloudify for Azure from first-generation clouds that required the development team to adopt the cloud provider’s language of choice. That can be a difficult aspect of what has come to be known as Platform-as-a-Service (PaaS).
”People get the direction to move to the cloud. Then they find out it can be much harder than they thought,” said Paul Burns, Analyst, Neovise. ”If they could move their applications ‘as-is,’ it is not that hard after all.” ”There has been evolution in thinking about the cloud platform as a service over the last year,” said Burns. ”That is why I see GigaSpaces [for Azure] as interesting. It allows developers to keep doing what they have been doing. Cloudify for Azure helps them move those apps out to the cloud.” Meanwhile, said Burns, the software takes advantage of the Azure’s elasticity, or auto-scaling qualities.
GigaSpaces worked with the Azure platform to ensure Java developers going ‘cloudward’ would have easier deployment, too. They ”would not have to implement [for example] scalability or high-availability on their own,” said Uri Cohen, VP of Product Management, GigaSpaces.
”It is important to say that Azure was built to run .NET,” said Cohen. “If you want something that is not .NET-based, you have to do the configuration, dynamic scaling, restart of failed components and [similar tasks] yourself.”
He said the new tools also provide better management of applications once they are up and running, adding that Cloudify for Azure was now in private beta, with a public beta due at the end of this month, and general availability anticipated by the end of the year.
GigaSpaces has evolved quite a bit since its first days. The evolution has led pretty directly to clustered cloud computing architecture.
”At first the company primarily focused on data caching,” said Massimo Pazzini, Analyst, Gartner. ”But it kept layering capabilities on the platform until, fundamentally, it became an application server – a container for building application logic and deploying this on a distributed grid of servers sharing the state of the application to the memory data grid.”
”Given that architecture, it was very natural for them to move into the cloud, which is a big, huge cluster,” he said.
The work with Azure is important to both GigaSpaces and Microsoft, Gartner’s Pazzini added. Much of the workload of future clouds will run on Java. And a cloud provider, be it Microsoft or Acme, has to sell CPU cycles.
“Microsoft has a huge infrastructure in place for Azure, and their problem is to sell the infrastructure as much as possible. They want to attract as much workload as possible,” he said.