Should we factor in quality of life when making decisions about data center operations? While the traditional metrics (resource utilization, operating costs, service levels, etc.) reveal the most critical information, do they tell the whole story?
I recently spoke with Tom Peacock, a data center manager at New England Biolabs. He introduced quality of life into the conversation and it got me thinking. In a field where “nights and weekends” are considered normal work hours, quality of life is a fairly revolutionary concept. Solid logic and hard economics normally carry the day, so looking at the soft side of life feels odd.
But when I thought about it, I realized there’s both a hard side and a soft side to this question. The soft costs of extra hours are obvious: stress, lost sleep, time away from family. But there are hard costs as well.
Beyond the overtime costs for hourly staff, there’s a built-in cost of inefficiency. After all, why are people working nights and weekends? Whether it’s to fix problems, or avoid downtime during peak hours, or just to get a project done, there are hard costs involved. Because implied in all of these extra hours is operational inefficiency.
Some tasks, such as adding an I/O card to a server, require downtime. Others, such as adding switches or routing cables, may take a long time to complete. Either way, you’re paying for it with less productive users and a tied-up IT staff. But it doesn’t have to be this way.
Server virtualization provides a perfect example. Before virtualization, moving an application from one server to another was a nights and weekends job. Now it can be completed in a few minutes. Server virtualization boosted resource utilization and operational efficiency.
Tom saw exactly the same potential with virtual I/O. By allowing I/O to be managed in real time and on live servers, he could improve quality of life around another set of tasks. He had a specific example from his experience:
“We had a connectivity problem with an ESX host. A new application required a private network but we had no extra server ports available. We had been working for a month to get the needed cards and cabling installed. With virtual I/O, we had the extra I/O configured in five minutes.
Plus, I have the comfort of knowing that if a server goes down I can transfer that entire VMware instance to another server by simply moving the I/O. Since there’s no remapping or reconfiguration required, I would be back up and running in minutes, not days.” (You can read more about Tom’s virtual I/O deployment here.)
He saved on hard costs by getting a job completed more quickly. And he’s saving on soft costs with reduced stress and the knowledge that if an issue does arise, he can work through it quickly… no extra hours required.
And the quality of life score is further enhanced in deployments where travel time to the data center is a concern, since virtual I/O let’s you address connectivity issues without actually being there.
The problem is, how do you factor this into your decision making? Some might say that soft costs cannot be accurately predicted and quantified, and therefore cannot be factored in. But there’s no doubt they become part of the benefit that’s ultimately achieved. So what’s your view? Do you factor in quality of life to your operational decisions, or is it all hard metrics?
Tags: Cost savings