Designing the modern data center begins with the careful placement of “good bones.” Data center network architecture must be highly adaptive, as managers must essentially predict the future in order to create physical spaces that accommodate rapidly evolving tech. Those with the best foresight on trends (including AI, multicloud, edge computing, and digital transformation) are the most successful.
As a consequence, all infrastructure management tools and processes are coming under close scrutiny by data center management, to ensure budgets are spent wisely. Optimization of design is a tough task. Previously, we talked about the basics of server rack design and overall layout and implementation, including cooling, power, and flooring structures.
For the third piece in this series, we address data center architecture on a higher level and look at how today’s leading subject-matter experts in IT infrastructure are making it work. It’s a common misconception that the growing popularity of edge networks and multi-tenant sites is somehow killing the data center. In fact, we’re seeing the opposite.
According to the AFCOM State of the DC 2019, these are currently the most sought-after cloud competencies:
- 53 percent — Data center, cloud, colocation connectivity expertise
- 41 percent — Cloud architects
- 40 percent — Cloud security professionals
Many experts in the field foresee continued reliance on the data center. Two of them have contributed to this series of articles: Chris Boyll, TAC Manager at Flexential, and A.D. Robison, VP Data Center and IT Services, North America at Rahi Systems. Both run strategy and implementation for major colocation facility chains—and colocation is one of the primary methods for creating agile growth as described below.
Within that same Gartner report, 62 percent of respondents say they are pursuing an initiative or program to make their business more digital. IT professionals are finding that they—and their company’s business managers—are more determined than ever to use the latest digital technology to gain a competitive advantage. Kaustubh Das, Vice President, Product Management and Strategy for the Computing Systems Product Group at Cisco, says: “In today’s digital world, IT needs to operate with a vastly different and more demanding set of user expectations.”
However, cost still remains king when making decisions about potential digital advancements. Here are some of the most popular (and affordable) methods for optimizing data center architecture and design:
Upgrading An Existing Data Center
Going all-new in terms of equipment isn’t reasonable for most DC budgets. Instead, IT managers blend the current tools with any upgrades deemed absolutely necessary. When these decisions involve racks and cabinets, heavy modification is not required.
“There is always a better mousetrap out there, and I’m sure somebody has thought of a better way to do things, but all of the racks that I’ve seen are very similar and haven’t changed much in recent years. You get different amounts of pass-through space between them, different ways to ground them and mount them, but most are very similar,” says Chris Boyll. Instead, the focus tends to shift to heating, cooling, redirecting the air, and replacing antiquated servers.
Co-Location or a Third-Party Facility
Businesses can benefit from the option of a co-location facility when tech is being developed more quickly than the infrastructure that was originally built to support it. This decision is an easy one at the end of a DC’s functional lifespan, or at a time of rapid expansion. Rack structure and layout in a colo facility is, by design, highly adaptable.
Robison notes that they never want the rack to be the limiting factor. “We’d rather have the limiting factor be what they paid the most money for—and that’s the electrical structure, first costs for mechanical cooling . . . we’d hate for customers to have to pay for that all over again because the technology has changed to the point where it’s no longer adequate.”
Colo managers are also accustomed to working with clients of all sizes and varying resources. “One customer sent us their equipment, and it didn’t have rails. We found an option that worked and set up the rails ourselves. We do what we have to do, to make it happen,” says Boyll. Rack integration also takes place off-site with regularity. Clients send cabinets fully cabled, tested, and ready for installation.
Go inside Flexential’s co-lo facility with Chris Boyll to see how ServerLIFT assists with day-to-day operations.
Multicloud and Hybrid Cloud
When moving to the cloud first became popular, some analysts predicted that commonly available cloud solutions would bring about the death of the data center. This proved to be untrue. Instead, a whole new generation of highly skilled IT professionals grew into new roles.
Many have adopted a hybrid approach. Hybrid multicloud is a combination of on-premises, public, and private cloud services. 93 percent of companies are looking to adopt the use of more than one cloud for their services.
The cloud complements and supports traditional data centers, as opposed to replacing them. Racks and cabinets now must support equipment that is even more expensive, dense, specialized, and heavy.
A move to incorporate cloud services frequently includes an in-depth assessment of existing software, apps, and platforms. Up to 50 percent of current infrastructure may be deemed ready for retirement. A data center manager tasked with adopting a multicloud strategy will therefore likely find themselves facing a massive physical transformation, migration, or consolidation of existing racks.
Going to the Edge
The digital expansion dynamic—the addition of intelligence to all areas of life—has had a significant effect on a number of industries, including retail, medical, industrial, and transportation. More and more intelligence is being pushed to the “edge” of the network, to the actual end device. The edge is outside both the cloud and the data center. It is the creation of new internet “endpoints” in decentralized nano-data centers. This “micro-data center” approach creates new opportunities and challenges.
Many edge data centers are remote, so they must run efficiently without a lot of maintenance. In edge DCs, server racks are modified in the following ways:
- Plug-and-play Full rack systems help to maintain speed and efficiency for the enterprise.
- Operators must build racks to hold dense, multilayer equipment supporting large amounts of data in a very small space.
- Racks must travel well, often over long distances, with rugged construction to withstand international transportation.
Build Your Own Data Center
This is arguably the most expensive of the options listed in this article, but it is not an uncommon choice for businesses concerned about security or equipment sensitivity. “15 years ago we could build a data center for $3 million per megawatt, [and today] that is just out the window . . . even with the most rudimentary systems, says Robison. “You’re looking at between five and ten million dollars per megawatt to build a fairly simplistic data center footprint these days.”
New storage and servers must be organized according to industry standards and whatever is required for optimal airflow, power, cooling, access, and connectivity. The equipment renewal process is undertaken every two to three years.
Design Implications for the Modern Data Center
“There is no perfect solution, in my mind, that accommodates everybody’s needs. We always need to be open to the possibility that something needs to change. If the change makes sense, we can always make it a standard feature. Or, it stays a one-off solution.”
– A.D. Robison
“For many of our clients, the need for a data center isn’t changing. What they need is a data center space that mirrors their aggressive and strategic digital growth.”
– Chris Boyll
No matter which modern design strategy is chosen, two core questions remain the same.
- How do you manage differentiated and geographically spread-out compute, storage, and networks?
- How do you do it safely?
This is as much a physical challenge as it is a data challenge. It’s helpful, therefore, to have all operational tools and protocols up to date. This includes guidelines on the physical handling of server equipment. These can be adapted along with the growth and development of your facility/infrastructure.
ServerLIFT offers internationally accessible equipment and tech support. Our data center lifts safely transport, position, and install and remove IT equipment for major transitions and minor moves. “Good bones” can adequately support the explosion of data seen in your industry. To continue the conversation, please share this article on social media with the hashtag #datacentersafety.