I recently attended the VMWorld Virtualization show in San Francisco. I’m sure I’m not alone in thinking, “This is not new.” Virtualization has been around for almost 40 years: IBM developed it on mainframes back in the late '60s. In those days of “the Glass House," uptime was a critical focus for the system operators, and mainframes were built differently from Intel/Windows boxes. Now that people are looking at going back to multiple virtual machines on a single physical machine, who is worrying about what happens if that physical machine crashes?
![Stratus Technologies [logo]](http://www.stratus.com/images/logo-on-white.jpg){kind=logo}
Is Stratus shipping ESX on the FTserver yet? I recall that it was coming the first of the year 08, was wondering if it was available yet?
General availability for VMware Infrastructure 3 (ESX) is planned for January 25, 2008.
There are many high availability solutions out there, it depends on what you need and if you have the skills in-house. Some fail-over solutions take more time then others to restart a server and/or a VM, this may or may not be a problem for you. I can see where a fault tolerant server might be a good fit here, but how simple is it to implement? Does it just ‘plug’ into an existing environment? Or is there something special that needs to be done?
That’s actually a great question ("is there something special that needs to be done?"). The simple answer is no! A Stratus ftServer system can very easily be ‘inserted’ into your current infrastructure and it will look like any other x86 based server from any other vendor. There is nothing “special” that needs to take place, no scripting, no special skills. From the operating system(s), applications, and in this case the hypervisor perspective, this is an Intel server. The big difference of course is the 100% redundancy and greater the five-nines (99.999+) availability. Also, what’s different is there is no ‘fail-over’ like there is with a cluster. With an ftServer, you get continuous availability, no loss of data, no downtime, no interruptions, no re-boots.
Do you think this would be a good fit for your virtual servers running your business critical applications?
That seems to be a major concern with a lot of people I’ve spoken to. With so many apps on a single server, if that server goes down the business impact will be huge. It’s not really one mission-critical application, but the combination of several that make it business critical. And apps that were not so important even just a few years ago, like e-mail, are business critical today. In many ways a lot has changed since the “glass house” days, and in other ways it has come full circle. Virtualization is one of the things that have come full circle — unfortunately the average x86 server is nowhere near as reliable as an old mainframe.