A dense data center does not inherently mean that it will have greater efficiency (typically, quite the opposite is true). However, implementing increased efficiency measures is necessary to the operation of these facilities as the trend of environments increasing in density continues across the industry.
Multi-cloud, public cloud, and private cloud are terms readily tossed around as the data processing capabilities and scale available to businesses needing to expand quickly through any of these cloud options continues to grow. As a result, the shift of data into the cloud has changed traditional data center operations and the cloud trend continues to drive the industry. The rapidity with which this trend moves means that concerns about the efficiency of a data center (especially when running cloud based applications) aren’t prioritized. Data centers and colos that are keeping up with the pace of the development of the cloud have to be able to scale quickly, and it is here that efficient operations – while still primarily ignored – make all the difference.
The leading challenge that the move to cloud computing has brought to data centers is the increased density any given facility is expected to handle. As servers improve, an increased data load can be run through a data center without compromising the actual infrastructure. However, scaling an environment to be this dense can have its consequences if the environment isn’t prepared. Geist recently published a blog on the rise of rack kw and comments on the pressure placed on facility managers to handle the rate of data being processed, saying “This is especially true for cloud and colocation providers, whose clients expect them to offer scalability.” Whether the manager is prepared or the facility was built for it, these data centers must rise to new challenges and increase data loads very quickly.
The “mammoth” providers leading this charge are inarguably Amazon Web Services and Google Cloud. Administers of public clouds, these companies are viewed with some concern as their dominance in the industry could allow them to completely take over cloud development. This has been proven false thus far, as various cloud options (public and otherwise) are being created with the ability to scale for certain cost tiers so that a multitude of companies have this option available to them, rather than hosting and running their own data center space. The CEO of Rackspace, Taylor Rhodes, was recently interviewed by ZDNet on his company’s role in shaping the cloud and what’s to come. Rhodes helps to position the reality of a move to the cloud, saying:
"When you move to a public cloud, really what you've replaced is your data center, your servers and storage and networking, the software that makes all of that work. But what you haven't replaced is IT operations management. Somebody still has to make sure that it's working and monitored, somebody has to do capacity planning, understand what you are going to spend and what should you spend next…”
It is this decision making that matters to the efficiency of an operation. The IT manager making these crucial choices already balances a variety of priorities, the utmost being that your data is maintained and running 24/7 with no downtime. However, upkeeping efficiency practices and solutions in the data center environment makes this job easier; it shouldn’t add an additional burden to an already overloaded workload for the data center manager.
Load balancing and auto scaling are existing opportunities to ensure uptime in a cloud data center or colo. Other solutions, such as a hyperscale rear door heat exchanger, offer additional opportunities for increased efficiency and options on how to better manage the data load being pushed through today’s enterprise and leading colocation facilities. Additionally, full containment, sensors to connect equipment that wasn’t built to be “smart,” and efficient best practices will allow the physical infrastructure to handle the loads required of servers. However, without an efficiency plan in place a facility will lose energy, lose cooling, and ultimately lose capacity.
Like an office building or any leased space, colocation customers come and go. As companies get bought or their needs change, infrastructure in colos or data centers is ripped out and replaced. Regular re-builds within any given space and changes in the whitespace environment require a re-instrumentation. Usually, this is expensive, so any capacity for energy savings should be capitalized on when a change in tenant occurs. Right now, times are good for the colocation market. Companies are less concerned with OpEx and are working to grow and meet cloud demand. As this slows however, they’ll start to recognize the places where they are going to have to be more efficient in response to the increased density that all data centers will experience in the near future.