The Evolution of the Data Center: What a Difference 17 Years Make

When I started my company, COLOTRAQ, back in 1999, the data center landscape was vastly different than it is today. The majority of service providers had come from the world of Commercial Real Estate and Construction and that culture had permeated much of how things were done. Back then, it was primarily about selling space with associated power and bandwidth or “power, pipe and ping” as we used to call it. Even the way the providers billed the customers and compensated the channel partners was indicative of that heritage. Customers were charged for the square footage their racks took up in the data center. The agents who brought the deal in were paid an upfront commission based on the total contract value (TCV).

Over time, many of these types of data center companies started to offer their clients managed services. The industry was quickly becoming commoditized. By moving further up the value chain (to managed services) data center operators were able to address the tightening of their profit margins. It was also a way for them to differentiate themselves from the competition. But many data center providers remained strictly entrenched in offering pure colocation and in some cases raw space. This segment grew to what we call today the “wholesale colocation” providers.  

Along came the high density server environments, several data center providers found themselves losing out on much of the revenue growth as clients were able to do much more in the same amount of space in their data centers. So in order to capitalize on this trend, these data center providers changed their billing model from charging for space taken up in their data center to charging the customers based on their power consumption. The argument there is that clients who were using high density blade servers were consuming far more power per square foot than clients with a traditional server footprint.

The other segment of the data center providers continued to evolve beyond simply offering managed services for their colocation clients. Some of these “retail colocation” providers began offering full blown managed hosting, whereby the client didn’t have to be bothered with sourcing and maintaining their servers and equipment at all. This group of service providers became very appealing to the mid-market sector who either didn’t have the internal resources required to manage raw data center space and equipment or simply wanted to focus on their core competencies and outsource the IT infrastructure management that supports their main business. This thought process was the genesis for the rise of cloud computing.

Cloud computing as a concept is truly a disruptive technology because it wasn’t driven by demand. It really began as some of the largest owners of data center infrastructure like Amazon realized they had all of this computing power that they really needed most of it during their cyclical retail seasons and specific times during the day. So they figured by selling excess and idle computing power, it would maximize the revenue potential of their data center footprint that they have already committed to pay for the space, power and bandwidth. This concept became wildly successful and ended up creating the demand we see today for this type of deployment. Clients no longer had to commit to a fixed term contract for space or power and have the flexibility of turning up and down data center infrastructure on demand. For the SMB market this was a no brainer. They have been the biggest and fastest adopters of cloud computing as they lacked the human and capital resources to manage their own data center infrastructure. Today, we see more adoption by mid-market companies; larger enterprises have been a bit more complicated.

While security and control remain major deterrents for the major enterprises to completely migrate their entire IT infrastructure to the cloud, many have either taken a hybrid cloud strategy or deployed certain non-mission critical applications on the cloud throughout their various business units. While many experts began touting that cloud computing will be the end of colocation and traditional data center deployments, nothing could have been further from the truth. What this cloud computing revolution has done is raise the whole ocean for data center infrastructure. Many companies are pulling the trigger on initiatives that in the past would take a back burner because of the commitments involved with traditional data center contracts. But in today’s world, IT organizations throughout all industries are able take advantage of this new computing environment. They can keep their most sacred mission critical infrastructure in a more traditional data center environment, while at the same time deploying various other applications, including pure concepts, in a cloud environment.

I’ve often compared the world of cloud computing to the auto industry. There was a time when the only way to get into a car was to buy it outright. Then chattel mortgages came about and many more people were able to drive a car by financing it. Then thanks to the growth of leasing, even more people were able to get into a car. And now with the advent of ZipCar and Uber, which I like to refer to as the cloud of the auto industry, even more people can use a car on demand and not be committed to using nor owning it for years. If you view mileage you travel as computing cycles, these business models are not just different ways to deliver the same service but really just a different way to bill the client for that service. The auto industry will not be suffering from this evolution as more cars will be needed along with parts and equipment and a growing civil infrastructure of roads and highways to support that growth. Certainly, no one is getting rid of their cars because they now have access to Uber and ZipCar. Quite the contrary it has generated more demand for auto transport and higher auto sales since Uber started making its mark in 2010.

At the end of the day, while there will certainly be a slight shift in demand for traditional data center services, the rubber has to meet the road at some point. Since the cloud lives in the data center, many of COLOTRAQ’s largest projects in recent years have come from cloud providers who are expanding their physical data center footprint. But the growth from their new segment of colocation users has far outpaced the decline in traditional data center projects driven by the mid-market and large enterprise market. As for the small businesses, we have definitely seen all of the 1U/2U projects disappear as these low value, high volume, high churn projects have gone to the public cloud.  

Over the last 17 years, having built the largest master agency focused purely on data center infrastructure (colocation, managed hosting and cloud), I can tell you that we and our channel partners have not missed those types of projects at all.