Manage Learn to apply best practices and optimize your operations.

Using metrics to fuel your data center ROI

If you need the upgrade you want, getting the right metrics is critical, Patrick Hubbard explains.

Patrick Hubbard, head geek and senior technical product marketing manager at SolarWinds, had a moving experience as his employer relocated its offices to another building. For IT, the move meant much more than just reconnecting a bunch of wires and cables.

In the span of a weekend, my company moved corporate headquarters to a new building. In the process, I witnessed an excellent example of IT admin dedication: our senior database architect fall off a pickup tailgate while moving an SQL cluster and its storage area network -- while every other company employee enjoyed a company event at a baseball game.

When your business transitions category, either from small to medium or from medium to large, IT admins are expected to up their game, often in unexpected ways.

The truth is, dedicated IT is a blood, sweat and tears endeavor. Management sometimes overlooks the fact that our recommendations usually come from hard-won expertise.

When we say "No," or more confusingly to some managers, insist on changing something, there is usually a good reason behind it. What's more, those recommendations don't only come from all-night config-a-thons, fat-fingering finance logical unit numbers or secretly pondering Boyle's Law while discussing cooling with the HVAC contractor; they also come from dropping batteries on our toes, losing the occasional nail on a speed rail, or playing find-the-glass in your finger after a shaky fiber termination.

And in our recent move, IT had the chance to demonstrate it has learned a lot, especially when it comes to using actual metrics to move budget managers and their demand for a data center ROI.

Changing business categories

When your business transitions category, either from small to medium or from medium to large, IT admins are expected to up their game, often in unexpected ways. For example, you might have been buying your patch cables from the same provider for years as your homegrown infrastructure scaled out over time. But what happens if it's time to completely rewire your data center? If you seize the opportunity to change to a new cable type, how will you test the replacement sufficiently in advance before ordering thousands of patch cables? It's a big component of the budget, and there's no easy do-over if you discover months later that you chose poorly.

You'll also be expected to plan new-to-you infrastructure, like advanced fire suppression, power and thermal management. To do that, you'll calculate business risk to weigh different options, plan N+1 cooling, consider business continuity and more. And, of course, the glorious future of expanded infrastructures (or at least remodeled ones) is still stuck in the budget ideas of yesteryear.

To get past the budgetary limits of your current sandbox, especially if your new sandbox is a real data center, admins need real numbers and not estimates to extrapolate the bigger spend.

Forget vendor ROI calculators: You're better than that

In our data center move, we increased the size of the center significantly, with impacts on every component of infrastructure. With cloud in the mix, I realized you need to know exactly how efficient your investment in every component actually is. It seems daunting at first -- after all, how can you calculate a single number that measures the service delivery cost effectiveness of the complex jumble of gear in your environment?

Turns out, it's actually not that difficult.

In our old data center, the admin team had been using composite monitoring to roll up the number for some time, and it was doing it with the management software it already had. All it took was a little elbow grease to make it happen. Admin members were using their virtualization management and monitoring solution to measure the real-world use of their virtualization infrastructure, including capacity, users, apps, hosts and spindles. From those combined metrics it's straightforward to project both the change in inventory and service delivery capacity.

They also used network performance monitoring to roll up the actual kilowatt usage from the rack power distribution unit-level, and related that to the gear in each rack. That produced power efficiency maps for every type of rack (and application) in the data center. Added to OS inventory, this returned licensing, power, networking and hardware costs to each delivered service level.

Last, they used their server and application monitoring software to measure server temperatures, fan speeds and other thermal metrics to map thermal load-to-service infrastructure.

Like rolling off a log (or tailgate)

Armed with this information and a couple of clever spreadsheets, the team completely removed speculation from the budget requests for the new data center. The team could plot growth across multiple dimensions, multiplied by the actual measured historical performance of their unique environment. They could also compare IaaS cloud options with on-site extension and consolidation alternatives on paper -- with charts and all -- in the lingua franca of budgetary decision making.

As a result, they got many new and shiny things.

And all of this was on the mind of our database architect as he and a senior network admin loaded the last database cluster server into the pickup truck. That must have been the distraction that led to the foot slip and the ensuing slow motion fall and skillful shoulder roll that saved his melon and the blade chassis simultaneously.

Resourceful to the end, administrators and architects make these saves every day-- usually invisibly. Next time, though, I might recommend a helmet to our database architect. After all, he has the master root password stored up there in his noggin.

About the author:
Patrick Hubbard is a 
head geek and senior technical product marketing manager at SolarWinds. With 20 years of technical expertise and IT customer perspective, his networking management experience includes work with campus, data center, storage networks, VoIP and virtualization, with a focus on application and service delivery in both Fortune 500 companies and startups in high tech, transportation, financial services and telecom industries. He can be reached at

This was last published in June 2014

Dig Deeper on Data Center Networking

Join the conversation

1 comment

Send me notifications when other members comment.

Please create a username to comment.

Have you altered your ROI assumptions when upgrading your data center?