* Has fault tolerance come across your virtualization radar?
I met with the people at NEC a few weeks ago to talk about the introduction of the company’s sub-$12,000 Intel-based fault tolerant server. The system is rolling out about a year after Stratus – a partner of NEC – began offering a sub-$10,000 fault tolerant system.
The interesting thing about these more affordable, highly reliable servers is that it opens up the option of fault tolerance to a wider range of customers – and to a wider range of uses. During our discussion, Michael Mitsch, a general manager at NEC, mentioned as an aside that NEC was hearing great interest from customers looking to use fault tolerant servers to harden virtual environments.
I found that quite interesting. A problem with virtualization is that you end up throwing all your eggs in one basket, so to speak. Yes, you’re making more efficient use of hardware by running multiple virtual machines on a single physical server. But what if that physical server fails?
To be fair, the virtualization software vendors do take that scenario into account and provide tools to move virtual machines among physical systems so that if there is a physical error a virtual workload can be shifted to a healthier machine.
But Mitsch’s comments got me thinking. As organizations put more critical business applications into virtual environments, they may find it worthwhile to invest in some fault tolerant systems to provide the physical foundation.
“Fault tolerant systems are really the only way to guarantee an SLA for a virtual container,” he says.
I’m curious, have fault tolerant systems come across your radar as you’re plotting out your virtualization plans? Let me know.
