Rack PDU Blog

Working as a data center IT professional is not unlike juggling chainsaws: there’s little room for error and a mistake could lead to disaster. While most data center IT pros are smart enough to stay away from flying power tools, they are tasked with juggling uptime, reliability, performance, and efficiency. And the stakes are high.

Bandwidth is the hottest commodity of the new year and will likely continue as such for the long haul. After all, both enterprise and consumer appetites will only become more insatiable as technology continues to require more support from the Internet. For example, enterprises are increasingly using the public cloud to run applications and store data due to its cost efficiency and simplicity. However, the more enterprises that utilize the cloud, the more bandwidth that is required to support their needs. Additionally, a surge in mobile video consumption is also impacting bandwidth requirements. This particular driver of bandwidth demand will likely multiply over the years as consumers continue to watch videos on social media, stream movies and use video chatting applications such as Skype or FaceTime.

We recently sat down with two of our in-house power strategy pros to talk about what the Server Technology slogan “The Power Strategy Experts” really means. The conversation helped to underscore the importance of doing one thing and doing it well, an idea that’s at the core of the Server Technology philosophy.

The Super Bowl continues to be one of the most watched television events in history, year after year. In fact, last year’s game where the New England Patriots took on the Seattle Seahawks became the No. 1 most watched televised event ever with an astounding 114.4 million viewers. And while you might be thinking about the outrageous strain on bandwidth—and thereby power consumption—required to broadcast this live event to televisions across the nation, the Super Bowl is actually one of the most energy efficient live broadcasted events.

Your servers generate a ton of heat over the course of a single day. So, why not take this excess heat and pump it back into your office building during cold winter months? It’s a great way to save money and reduce waste at the same time. Think about it: A typical server running at 400 watts will generate almost 1,400 British thermal unit (Btu) per hour. A standard 20-by-20 room with a 10-foot ceiling only needs about 35,000 Btu per hour to maintain a 72 degree temperature. Since a large enterprise will run several thousand servers at a given time, it will generate more than enough energy to keep an entire building heated for an extended duration.

What you’re looking at is a long-term forecast of heat load by footprint, specifically made for the IT datacenter and the telecom market. Today, the upper limit of air-cooled compute infrastructures sits right at 4.5kW / ft^2, which aligns nicely with ASHRAE’s prediction. Call it a self-fulfilling prophecy or call it an advancement in cooling and liquid immersion systems, the end result is a slow, but steady rise in compute densities over time.

In a recent post, we ruminated over the fatal design flaw in the Death Star’s thermal exhaust port that ultimately led to the destruction of the space station. As we mentioned, it’s difficult to pin blame on the Death Star architect alone, as all of that exhaust had to go somewhere...but, we digress. Check out the article and share your thoughts with us. Fast forward to The Force Awakens: The Galactic Empire has fallen, and the First Order has commandeered an abandoned ice planet in the Unknown Regions. Enter Starkiller Base, a massive facility capable of destroying not just planets, but entire star systems. Now, you’d think in the First Order’s quest to establish itself as the new head honchos in the galaxy that they’d at least not make the same mistake that the Empire did and leave their fancy new weapon vulnerable to attack. But, they did.

There is an array of variables that affect your data center uptime — from structural design to redundancy to fault tolerance. You could say that it’s a balancing act, ensuring that every element positively influences the others so that damaging issues —like power shortages — don’t tip the scale.

If you gathered data center operators from some of the world’s top brands to discuss capacity planning, you could probably pick up a helpful tip or two on how to plan for a new build or retrofit. In fact, that’s what our friends at Datacenter Dynamics Intelligence did when they set out to interview 15 data center owners and operators from companies as varied as Cisco WebEx, Defense.net, and ING Bank about capacity planning best practices. The resulting study provides a wealth of common-sense tips for those tackling either new builds or retrofits.

here’s no crystal ball to predict the future of your data center’s power density, but if you have a clear understanding of your data center’s goals, you can cut through much of the confusion that revolves around density while being prepared for eventual increases. In recent years, high density seems to have become a relative term with various data center operators having their own ideas about what it really means. For a long time, the data center was an amalgam of old mainframes, networking racks and a variety of server/storage cabinets. Around this period – roughly 15 years ago – most racks were running the 1-3 KW range with power and cooling infrastructures being overbuilt to attain maximum uptime. By the time 2005 rolled around, data center managers were seeing increased rack power, especially on blade servers.