Andres Rodriguez - Fotolia
As IT teams take on a more strategic role in the enterprise, the challenges they face become tougher.
Ranging from tricky outsourcing decisions to mainframe issues, here are some of the biggest data center challenges our advisory board members faced, or saw others face, in 2016, along with recommendations for overcoming them in 2017 and beyond.
Clive Longbottom, co-founder, Quocirca
Last year was the least predictable year that many data center managers have seen in their lifetimes. The good ol' days of planned expansion based on five-, seven- or 10-year facility lifecycles has disappeared. Now, few can predict whether they will even have a data center facility in five years -- never mind what size it will be, or what technology will be in it. The role of the data center manager changed from one of facility planner and operator to advisor, helping determine where to host business workloads -- an owned facility, a colocation facility, or a private or public cloud.
However, at a more prosaic level, data center managers must still maintain existing facilities with the best possible costs and performance. Power efficiency at the equipment level, combined with the best cooling at the lowest, ongoing cost, has remained a focus, although unpredictability has also had an impact here. The total lifetime value of a change to cooling is almost impossible to figure out. Previously, managers could assess this over a relatively long time, but now it has to be over no more than a three-year period, given the likelihood that the facility will have to change significantly after that. However, as newer technology at the platform level comes into play, such as hyper-converged systems and flash-based storage that can operate at higher temperatures with more targeted cooling, it still makes sense to review the use of existing cooling systems during 2017.
Robert Crawford, systems programmer
I spent a lot of this year dealing with technical debt. My [mainframe] installation, like many others, has a lot of utilities, customization and glue code written in the 80s and 90s. This code provided functionality unavailable at the time and improved performance.
Thirty years later, the motivation for these enhancements has evaporated. Much of the functionality supported by customization is now available in system code. The internet age tends to prioritize flexibility over performance. Lastly, a lot of the original coders have retired and few people understand assembler.
As a result, we have problems upgrading our applications or implementing new techniques due to incompatibilities with the past. Basically, a big bang approach is too risky and an evolutionary scheme would be too time-consuming. Luckily, my shop recognizes the problem and is committing time and money to make a road forward to overcome these data center challenges.
By the way, these issues won't be unique to mainframes. Some web applications are more than 20 years old and, for all intents and purposes, are legacy code. I'm sure in the next few years, when someone wants to make big changes, they'll find all sorts of shortcuts, customized code and security hacks. At that point, they'll repeat my mantra: "I guess it seemed like a good idea at the time."
Lex Coors, chief data center technology and engineering officer, Interxion
To give you an overview of the industry's biggest data center challenges in 2016, I first have to go back in time. Six years ago, I was asked how long the [colocation] industry would continue to grow at this speed, and my response was six years. A year later, we started to realize the role that the cloud was going to play in propelling the industry, as users increasingly looked to colocation to house their most critical applications and leverage the carrier-dense networks. And, in just the past couple of years, as we've started to see the benefits of IoT and artificial intelligence, we can only imagine what this means for our industry in the coming decade.
The biggest challenges for the colocation industry in 2016 were to ensure that we're enabling organizations to digitally transform their business by offering them access to hybrid cloud environments, and continuing to adapt to constantly changing IT demands.
Explore predictions for data center trends in 2017
Quiz yourself on software-defined trends
SSD dominates the market and triggers a transition