The traditional data center of yesteryear has been replaced with lean, mean virtual machines designed to keep pace with the insatiable demand for compute and storage resources. The environment has changed so dramatically that certain parts of the data center may no longer be recognizable—racks of name-brand servers, storage and networking gear are being jettisoned in favor of open-source solutions and white-box platforms that promise major cost savings and unprecedented flexibility.
Because of all this change—or perhaps as a result of it—the number of data centers continues to shrink. According to IDC, there are fewer data centers owing to consolidations, the cloud and the trend to rent IT infrastructure. As a result, new economic opportunities are opening for companies that need to quickly build and scale robust data center architectures. This approach, however, also can lead to undue risk and unforeseen support costs.
That’s why it’s critical to embrace the changing data center landscape with an open mind. Instead of focusing on what has always been done, look to emerging directions, such as open computing environments and more-flexible support structures. Without losing sight of what matters most, what should change are some of the traditional routines and methods regarding maintenance and managing equipment life cycles. What follows are four areas for consideration when tackling data center transformations.
Hybrid Clouds Deliver Greater Benefits, Less Risk
Moving to the public cloud makes great sense, but not every app or service should reside there. Among the early adopters, SMBs typically gain a big uptick in capabilities by moving to the cloud because they typically lack the financial and technical resources to deploy and maintain robust infrastructures on their own.
The larger the company, however, the less it can gain from an all-in public-cloud policy. Too much is at stake to put everything in the cloud, as all it takes is one service outage to disrupt business and cause customers to lose confidence and trust. For that reason alone, it’s safe to say no more than 50 percent of critical business workloads should reside in a public cloud.
Privacy and security, especially in regulated industries with PCI and HIPAA requirements, remain too important. They’re always on the list of the top five things CIOs are thinking about. That said, although the public cloud definitely has a purpose, hedge your bets because we have yet to figure out exactly when the pendulum reverses.
Brand Loyalty Is a Thing of the Past
Ten years ago, everyone had a homogenous data center: all HP or IBM. Five years ago, things were becoming heterogeneous with an ever increasing mix of Cisco UCS and a smattering of HP, Dell and IBM. Today, brand-name products have been usurped by a wave of white boxes that dominate many data center environments.
As such, there’s little to no brand loyalty—perhaps with the exception of Cisco, which has done an excellent job of sustaining the perceived worth of its various certifications. Instead, data center decisions today are driven more by price-performance considerations than the value of a particular product brand.
Unlike the data center mainstays, white-box vendors haven’t been around for 20 years or more, so they haven’t had time to build and sustain world-class services and support organizations. What this situation means for customers is they’re largely on their own when the box breaks.
Without saying it, white-box vendors expect customers to be happy enough with price savings and new flexibility that they’ll just buy a new box when they run into a problem. After all, they’ve done their part in unshackling companies from demanding name brands, even if the liberation comes with a few unforeseen risks. In particular, the growing number of low-cost, offshore options raises questions when it comes to identifying the level and quality of support that comes with each white box. Luckily, the growing availability and viability of comprehensive support choices from independent, third-party service providers can go a long way to mitigate reliability and support risks.
Reliable Support Is More Important Than Ever
The liberation that comes with increasingly heterogeneous equipment decisions and the growing popularity of white-box offerings makes finding reliable and cost-efficient support and maintenance a critical part of every data center equation.
Expect the continued emergence of white boxes to propel the need for growing ranks of third parties that provide a robust roster of support and maintenance options.
When evaluating these options, avoid being enticed by low-cost service organizations composed of “freelancers” and/or subcontractor support. The tradeoff here is that low cost and quality support seldom go hand in hand. This is a clear case of getting what you pay for. The best approach to ensuring consistent, quality support and expert professional services is to evaluate providers’ ability to offer rapid response and direct-delivery support options.
Alternative support organizations not only can fill the services void created by the proliferation of white boxes in data centers, they also enable companies to keep viable legacy gear up and running longer after vendor support expires. The long-held practice of using third-party maintenance to support network gear is becoming more prevalent in data centers as IT managers see the benefits in extending the useful life of time-tested and field-proven server and storage equipment.
Consider Extending Product Life Cycles With Pre-owned Equipment
For years, IT organizations in companies of every size, shape and industry have relied on pre-owned networking equipment in some manner to keep their networks running. Taking advantage of “the secondary market,” as it’s often called, to buy certified gear is one of the handiest arrows in any CIO’s quiver. After all, it’s hard to resist savings that can be as high as 90 percent off manufacturer list prices as well as the option to avoid being forced into the latest and greatest when all the new bells and whistles are unneeded.
The same can be said on the server side. Although the pre-owned market for low-end servers is pretty well known, the midrange- and legacy-product categories offer excellent alternatives to maximizing data center budgets. This approach plays well to the current trend in many data centers, which resist additional infrastructure investments in favor of extending the life cycle of existing equipment or repurposing it while eyeing what next-generation gear will look like.
Teamwork: How to Survive and Thrive
Surviving—and thriving—in the changing data center landscape will require blazing some new trails and wearing some new hats. Fortunately, it’s not a mission you must handle solo. Teamwork with clearly delineated roles are the answer. Rely on trusted colleagues in the data center for their in-house expertise while maximizing the capabilities of third parties for their specific talents.
About the Author
Mark Nichols is chief product officer for Curvature, an independent provider of IT network and data center life-cycle management and services.