The new Coolan TCO Model can help web host data centers lower costs and reduce downtime, ultimately passing those savings and benefits to customers. At the 2015 Open Compute Project (OCP) summit, OCP founder Amir Michael surprised everyone by springing an underdog subject to the crowd, which was “something not often talked about in the industry. There is a trend of people moving off the cloud.” In an age where everyone is talking about moving to the cloud, Michael’s assessment sprung confusion. However, he says the analytics platform Coolan doesn’t lie. Along with his brother Yoni Michael and friend Jonathan Heiliger, the trio founded Coolan to aid the owners of data centers in minimizing downtime and keeping costs as low as possible. Coolan analyzes server data to see performance levels and pinpoint problems. In early pilot testing, Michael saw that a lot of people didn’t know where to put their data storage or data processing.
“Trying to figure out the total cost of ownership (TCO) for your compute infrastructure is no easy task. With the proliferation of public, private, and hybrid clouds, not to mention the ever-growing number of Buzzword as a Service paradigms, anyone in charge of their company’s infrastructure faces a difficult decision: Should you build or acquire your own data center, lease in a collocated facility, or rent a piece of the cloud?” The answer to these questions may surprise you, but it’s nobody’s fault the truth wasn’t revealed sooner. After all, Coolan has only been around a short while.
Coolan the Gang
Michael is the first to know how challenging developing Coolan was. However, he’s always up for a challenge, and partnered up with Jimmy Clidaras, who’s a well-established engineer specializing in data centers and infrastructure at Google. The two had already been working on projects together when they agreed that TCO in the data center industry was astronomical—and may not need to be. In just three months, they created a TCO model that looked like a basic spreadsheet, but was actually steeped in extensive analysis. It took several research hours from a skilled team to come up with this 11-page spreadsheet model.
According to Michael, “We built the TCO Model to let operators evaluate and understand the cost points for different types of deployments.” The cost points are many, but some of the biggies include best hardware components, data center PUE, capital costs, how much power networks are consuming and whether or not a system upgrade is in order. Of course, the biggest question of all was: Cloud, colo or private data center?
And the ultimate question: stay in the cloud, go colo, or build a private data center?
Curiouser and Curiouser
At the summit, Michael addressed some of the most astonishing findings from his team. Surprisingly, power costs weren’t outlandish. According to Michael, electricity costs for a special, in-house data center came up to just three percent of overall costs. For a colocation center, it amounted to five percent. He’s also quick to point out that “standard methodology” isn’t always the best approach or even appropriate for every data center owner. Just like people, every data center is different.
In one example Michael gave, a fictional data center is best suited to managed services via the cloud because they needed less storage and more computing power. The actual storage amount needed makes a big difference when deciding on deployment options. Growth costs are less for in-house data centers compared to leasing at a colo site or cloud-managed options. In the end, Michael says he knows the findings fly in the face of today’s “standards,” but the numbers speak for themselves.
What Michael is really amped up about is the fact that customers are getting innovative, thinking outside the box and coming up with new ways to run data into various models. It’s Coolan’s hopes that the TCO model will be open-source, serving as a free platform to increase industry transparency. This will let data center owners see their costs clearly so they can make better decisions for their centers, their budgets and their customers. As more clients understand the differences in data center models, they’ll want more transparency, too. This TCO model is a great way to bridge disparities and streamline what can be a confusing decision.