Rack PDU Blog

It’s no secret that data centers are massive energy hogs. As explained in a recent Server Technology white paper, “The Power of Hyperscale Compute,” a typical data centers can be 10 to 100 times more energy-intensive than an office. And altogether, data centers use about 3 percent of the U.S. electricity supply. Over the last two decades, as compute and storage densities have increased, rack power densities have also skyrocketed. In the past, a typical rack would consume an average of 1 to 2 kilowatts of power. Now, as we move deeper into the hyperscale era, loads are hovering around 20 to 40 kilowatts. More servers and hard drives are being put into single racks today than ever before, in a scale out approach.

You have a deadline, and you have your goals. Your hyperscale data center design needs to maximize power efficiency. Use free air cooling or adiabatic cooling. And support ambient air operating temperature of 25-35oC, with hot aisle exhaust temperatures approaching 60oC. You need lots of outlets – C13s, C19s, or even something that allows you to blind mate servers to the power strip. And it needs to meet regulatory requirements in most major geographical regions around the world.

What is a PDU? This may seem like a rather existential question, particularly from the blog space of one of the leaders in the marketplace. So why ask such an obvious question? Well, because the answer is not entirely clear. The mission-critical data center world is chock full of acronyms, those pesky, foreshortened forms of communication that allow us to sum up so much in an efficient amount of space. But when they take on a life of their own, they become like the little monsters in the movie ‘Gremlins.’

I must admit I have been thinking a lot about quality lately. Our data center industry, at its core, is about power correction, conditioning, and quality. The ability of a mission-critical electrical system to deliver the proper sine wave to the server is a multi-billion-dollar industry. (My apologies in advance to the mechanical minded for such unilateral statements.)

For those of you who are working through the issues of remotely managing your data center power, or are thinking about heading in that direction, we have a few pearls of wisdom. There is definitely some confusion in the industry about smart versus switched power distribution units. I’d like to take a minute to talk about the latter. Although both types of rack PDU’s have the same features such as circuit protection, monitoring, security, and communications, it is the switched variety that opens the doors to remotely managing your data center.

At the beginning of the sunny season we talked about some things for school system and university data center managers to review during the summertime break. Seems like yesterday.

In this recent article on Data Center Knowledge, the author discusses the growing issue of data center real estate in Singapore and other APAC markets – or more accurately, the lack thereof. While many countries, including portions of the United States and Northern Europe, benefit from large landmasses, high-density city centers in most of the world place an absolute premium on every square meter of space. In these areas, data center facilities are built more on the vertical axis, with support equipment stacked on rooftops and in basements.

e were recently approached by a C-level executive who was facing a challenge. His company had recently acquired another organization, and was now in the process of trying to determine the best way to integrate its data center. The data center, however, was in need of a major overhaul. A big part of the problem was the fact that the former company’s IT team had been using traditional spreadsheets to track its devices and daily power metrics. These spreadsheets were rife with errors (and possible manipulation). For example, the team discovered several network devices in the data center that did not appear in any records.

Earlier this year there were many questions in the data center industry about what would happen to the Federal IT Acquisition Reform Act (FITARA), a 2015 law that is set to expire in 2018 and 2019. It’s taken a few months, but now we’re starting to see some answers. Just recently, news broke that there is a new version of FITARA being proposed in the House version of the 2018 National Defense Authorization Act. So FITARA does not appear to be going away anytime soon.

You asked, we answered. During our highly anticipated webinar last week, "Delivering Resilience and Uptime at the Edge," our speakers were asked a few questions in regards to the challenges presented by edge computing, their views on its future development, as well as how Server Technology's portfolio of products will address its demands.