With internal data center traffic expected to grow threefold over the next five years - and the penetration of Internet of Things (IOT) devices predicted to be 30 billion strong by 2020 - enterprise data center bandwidth is at risk of being overloaded in the near future. Same could be said of colocation facilities as well. These are hardly revelations anymore. Technology builds upon itself, so its growth is almost always exponential.
Needless to say, data center managers are having to adapt to some pretty tough challenges: applications demanding higher lane speeds and ultra low-latency performance; increasing port densities that can support leaf and spine networks; and, while you’re at it, find a way to improve network availability while lowering costs across the board.
For many data center managers, the answer lies in migrating their infrastructure - either with rip and replace or slowly over time—to support the new speed, latency and port density requirements. But is this both necessary and true for all data center facilities? Absolutely not.
On the one hand, there is no ignoring the fact that current and future data usage trends are alarmingly high and not expected to level off in the foreseeable future (read that as “ever”). So infrastructure migration is not a question of if but when. On the other hand, every data center facility has a very unique set of business requirements, stakeholder expectations and technical considerations.
Looking across the region, Equinix bought TelecityGroup for $3.6 billion, gaining more than 40 data centers in Europe, according to DCD. Meanwhile, Apple, Microsoft and Google are investing in Irish, Dutch and Scandinavian facilities. Big US providers and enterprises are fortifying their presence closer to their end users across the Atlantic while IBM recently opened two cloud data centers in London. Infrastructure migration, therefore, becomes a matter of timing. So, if not now, when is the right time?
Pay attention to the standards. One of the requirements for proposing development of a new standard is being to demonstrate the need for and feasibility of the new standard. In other words, bodies like IEEE, TIA and ISO typically will not initiate work on a proposed standard if there isn’t much of a market need. So if a standard is being developed you can bet there is sufficient market demand for whatever is being standardized.
Know your business. Data centers do not exist in a vacuum. Every facility plays a role in the greater ecosystem of the organization (or market, if you’re a colo). Understanding the degree to which the organization relies on the data center is critical. In some cases, prepare to be surprised. A 2015 study looked at the investment decisions of managers at over 1,000 companies and discovered that gut instinct tended to drive decisions rather than data.
What’s the cost of standing pat? If you’re going to delay your migration, make that decision with eyes wide open. New applications or sudden changes in market dynamics can increase the bandwidth requirements overnight. As data center demands grow, what’s the effect on your infrastructure and how will that impact manageability and potential downtime?
The bottom line is that your data center exists to serve a particular role. To determine if the time is right to migrate to higher speeds, ask yourself this simple question: “Is my physical layer holding me back?”
At CommScope, this is the stuff we think about every day. To help you stay focused and on-track, we recently teamed with DatacentreDynamics to produce 4 Steps to a Successful High-Speed Migration Strategy webinar. This webinar cuts through the noise for a practical top-down approach to help guide your migration strategy. Watch the replay of the webinar here. Then contact CommScope and learn how your vision plus our High Speed Migration platform can take your data centre beyond the now.
John Schmidt is vice president of global data center solutions at CommScope