Data Center Challenges: Top 3 Design Considerations
Prior to the pandemic, data center demands were relatively predictable. Not anymore.
With the remote worker locations constantly shifting and the further proliferation of connected devices, network management has become a major challenge. And it’s not just about provisioning enough routers and switches to keep the data flowing. Ensuring uptime, reliability, and security is far more complex in this highly-distributed environment.
Storage and compute needs are also harder to foresee. The ongoing tug-of-war between cloud-based, on-premise and hybrid architectures means you never know where your next application will reside. But there is little room for error, as data centers need enough speed to run today’s high-bandwidth, data-intensive applications—including video streaming and real-time AI.
Fulfilling these pressing, ever-evolving needs requires intensive research and planning. In the recent Design Engineer Tell-All survey commissioned by Molex and Digi-Key, more than 40% of designers said that higher customer expectations and increased design complexity are expanding their workloads.
At the same time, clients are demanding greater operational simplicity for the front-end employees, partners and customers who will use their infrastructure. They also want designers to contain burgeoning energy costs and get their new data centers up and running as soon as humanly possible.
Managing these competing demands in a pressure cooker is no easy task. By setting strategic priorities, evaluating options from a lens of client needs, and conducting thorough testing to ensure that systems perform to expectations, designers realize the best chance of success.
Here are some fundamentals to consider:
- Improving energy efficiency
Data centers are currently responsible for about 1% of global electricity demand, and video streaming, online gaming and emerging digital technologies are likely to boost demand even higher in the future, according to the International Energy Agency. Storing and managing more data means generating more heat and spending more on cooling the data center.
Location plays a significant role in determining these expenses. To ease their clients’ cost burdens, some designers are building centers in Arctic Circle countries such as Sweden and Iceland. But operating a data center far removed from business operations isn’t practical for everyone.
Fortunately, engineers have improved efficiency in many server power supplies, storage devices, network switches and other infrastructure components in recent years. Designers should consider the energy consumption of every piece of infrastructure they purchase, including small equipment such as connectors. When added up across the data center, even low per-unit costs can become significant, especially in today’s environment of rising electricity rates.
Clients’ energy needs are also changing as companies transfer more of their on-premise services to cloud hyperscalers. Designers should discuss clients’ digital transformation plans to assess their potential impact on data center energy efficiency.
- Optimizing connectivity
Surrounded by responsive consumer devices and applications, today’s employees and customers won’t tolerate slow or spotty connections. Designers must balance the need for speed against the cost of technologies such as fiber optic cable. Depending on client needs, less expensive methods of lowering signal loss, such as linear amplifiers/retimers or additional cabling, may be prove valuable.
Providing connection redundancy to ensure business continuity and disaster recovery is one of the design engineer’s most important responsibilities. They must also make sure redundant nodes can communicate with each other effectively.
Another important consideration is product integration. Because applications now have many points of interconnection, designers need to map data flows throughout boxes and systems to determine how to achieve the lowest possible latency.
That can be difficult to do when technology is produced by different manufacturers. In the Molex survey, 19% of design engineers said their greatest challenge was ensuring effective integration of multiple technologies. While it can be tempting to avoid integration problems and obtain favorable pricing by using a single vendor for multiple components, designers will serve their clients better by avoiding lock-ins and evaluating each piece of technology on its own merits before analyzing connectivity.
- Testing product effectiveness
Designers are under great pressure to bring their data centers to market quickly, but they should explain to clients that the cost of failure is greater than the cost of delay. For example, if a product such as a server switch is discovered to be defective after all the servers have been set up and plugged in, the labor costs of accessing and replacing it, coupled with the associated downtime, far outweigh the impacts and cost of the additional time it would have taken to run a test.
Establishing a safe and efficient data center requires designers to conduct comprehensive testing to ensure every piece of infrastructure meets quality control standards. In the Molex survey, 88% of designers said they had access to a lab or “maker space” stocked with parts and tools for experimentation.
By experimenting with variety of components, design engineers can select those that not only meet client requirements, but provide the highest level of performance, consistency, and flexibility. In addition to lowering service costs down the line, using high-quality, adaptable components can give clients a competitive edge by allowing them to adopt advanced capabilities sooner.
Designing a data center for an environment of changing business practices and advanced, interconnecting applications is no easy task. By selecting the right location and carefully analyzing infrastructure for energy efficiency and connectivity, designers will help their clients save money and avoid productivity-draining conflicts. And by doing comprehensive component testing, they can deliver durable, flexible, state-of-the-art equipment that allows companies to adapt as technology continues to evolve.
As a trusted advisor to some of the biggest and most well-known names in data center environments, Molex is embracing and adapting to the changes across the data center landscape today to ensure enterprises can achieve the most value tomorrow. Molex’s broad portfolio of next-generation connectivity solutions leverage the latest advances in copper and optics to deliver high signal integrity, lower latency and reduced insertion loss for optimal efficiency, speed and density.