Microsoft, Marathon Team On Windows Server, Hyper-V Fault Tolerance
The two companies said Friday they are jointly developing the integration between Marathon's everRun fault-tolerant software with Microsoft's Windows Server 2008 operating system and the Hyper-V server virtualization platform.
Fault-tolerant technology takes advantage of redundant components in a server to ensure that an application remains running even if some component in the system fails.
The goal is to add fault tolerance to business-critical applications which absolutely have to run 24/7, said Mike Schutz, director of product management in Microsoft's Windows Server division.
Microsoft currently offers a basic high-availability solution with its Microsoft Clustering software. However, when a server in a cluster fails, the other server or servers and their applications need to be started. Fault tolerance provides redundant computing so that if one server fails, the other continues to run without impacting the application.
The need for fault tolerance in Windows environments is growing, Schutz said.
"Microsoft is seeing more and more customers relying on Windows Server to host business-critical applications which obviously increases the need for reliability," he said. "Also, the growing impact of virtualization is putting the spotlight on reliability."
Marathon's everRun fault tolerant technology currently works with Windows Server 2003, but will also work with Windows Server 2008 when the R2 version becomes available in the second quarter, said Gary Phillips, president and CEO of Marathon.
It will also be ready to run with Hyper-V in about 12 months, Phillips said. It already runs with virtual servers configured with Citrix XenServer, but there are no plans to port the technology to VMware's ESX environment, he said.
everRun, combined with Windows Server 2008, offers customers three levels of protection which can be set in a granular fashion depending on the requirements of the virtual server or the workload, Phillips said.
The first is the basic Microsoft Clustering that already comes with Windows Server.
The second is component-level fault tolerance, which Phillips said mitigates failures due to the loss of some component in the server.
The third is system-level fault tolerance which protects the workload in the case of a failure of an entire server by mirroring two servers to work on the same workload, he said.
Marathon depends on the channel for about 100 percent of its revenue, Phillips said. The company has about 450 partners in 45 countries worldwide, he said.
Under the agreement signed between the two, Microsoft and Marathon will co-market their solution, Schutz said.