"Without being able to decide which workloads you're going to put where, and where the resources are that make you want to put one on a particular server, you can't move up the stack toward cloud or make your virtualized infrastructure run as predictably or efficiently as you should."
Just being able to see all a company's VMs – both where they are and what they're doing – is enough of a reason to spend money on management products according to Jon Shulda, systems administrator for Rotary International, which is based in Evanston, Ill.
"We're around 80 percent, 90 percent virtualized," Shulda said about Rotary's 300-server infrastructure. "As we became more virtualized we were losing a lot of visibility into what was happening inside the hosts until, like any other organization, frustrations about performance started being blamed on the thing we understood the least.
"So any time there was a lag we'd figure it had to be a bottleneck in the VMware servers, but even when that was right, it didn't tell us enough about what to do about it," Shulda said.
VMTurbo doesn't answer all those questions, but it answers a lot, and predicts thanges that may have to be made later, to improve capacity planning.
"It lets us look down the line and find where the max load will be on the environment, identify where the latency will be, and work around those things to keep costs down," Shulda said.
"I agree that the whole cloud concept is undefined, but I don't think that's what the organization is looking for from us," Shulda said. "We run [Vmware's clustering] Vblock so, for all intents and purposes, we run a cloud network, but that's not what we're after.
"What [the decision to try VMTurbo and VMware cloud management apps] amounted to was software makers getting to the point of supporting virtualization to the point that it makes more and more sense and works that much better and lets you see what can go wrong," Shulda said.
Doesn't sound like a miracle cure, as cloud often does. That approach is much more practical than metaphorical clouds.
But it works; it works now; it saves money and delivers better performance and divides computing resources according to the needs of the apps due to receive them.
There's a lot of that in descriptions of how the cloud works, though it manifestly is not as sexy or market-slick as the hottest buzzword-metaphor-tech ology on the market.
But it works a lot better.
And, by coincidence, it's almost the same thing, anyway.