Measure, measure, measure: It's one of the pieces of advice repeatedly doled out to companies looking to make their datacenters more energy-efficient. Without question, it's a sound suggestion; if you don't know where you're starting from, how do you know where you should go or whether you've made any progress?
The real difficulty, however, lies in the measuring process. It can be time consuming. And once you've thoroughly measured the temperature, humidity, airflow, and energy consumption throughout your facility, you have a valuable snapshot -- but just for a while. As time passes and changes occur within your datacenter -- new equipment is added, a tile plate comes loose, a CRAC unit goes down -- you might not realize you have a hot spot until the next time you take a measurement (or a system overheats).
Fortunately, change is afoot: Companies are rolling out smart tools to help datacenter operators measure key datacenter metrics in real time, thus giving them unprecedented insight as to how efficiently their facilities are running at any given moment. That's the long-term vision, anyway. Some, such as Microsoft, are doing it internally while others are bringing these systems to market.
One such intriguing offering is coming from a relatively young company called SynapSense. The company has devised an innovative solution that uses wireless sensors to gather various types of data on temperature, humidity, and energy consumption. The system software then crunches the numbers to provide status reports and information on datacenter health through a GUI. You can also set it up to send e-mail or SMS alerts under certain conditions, such as an abrupt drop or rise in temperature.
The most recent addition to the company's product line is called LiveImaging, which creates thermal, humidity, and pressure-differential maps of a datacenter using 3-D real-time data. Thus, an IT operator could log in to the SynapSoft console at any moment to find instances of hot spots or direct airflow, then take necessary actions to address overheating and air mixing.
SynapSense isn't the first company to use sensors to monitor the temperature in a datacenter. HP, for example, uses them in its Dynamic Smart Cooling offering. However, SynapSense differentiates itself with its wireless sensors. These minute devices run on battery power, which means you don't need to deal with plugging yet another device into the datacenter power strip.
Chuck McParland, a computer scientist at the Lawrence Berkeley National Labs, also sings the praises of the wireless approach. "Wireless sensors will have a huge impact on the way we see our environment -- from monitoring energy usage to controlling environmental pollution," he says. "Today, [SynapSense] is providing us with datacenter energy usage information that would simply be too difficult to acquire with any other technology."
SynapSense has a long and winding product road map ahead. For example, it's currently beta-testing tools that would enable a datacenter operator to view a facility's PUE or DCIE in real time. (PUE and DCIE are metrics created by The Green Grid to help organizations assess the overall energy efficiency of their datacenters, comparing the total amount of power a facility uses with how much of that power actually gets applied to computing instead of cooling, power conversion, and other support functions.)
Supporting SynapSense's ambitions, the company has strong backing from venture capitalists, having landed US$10 million in Series B funding last September. The company also has backing from IBM, which has implemented SynapSense's technology and declared it among its top startup partners.
Additionally, power utility SMUD (Sacramento Municipal Utility District) recently completed a successful pilot program running SynapSense's technology and is looking at an incentive program to get customers to deploy it. PG&E is doing the same.
SynapSense's product is an intriguing one, and other vendors, both well established and up-and-coming, will no doubt design and revise their own wares to provide similar -- or superior -- functionality. Although datacenter operators should already be measuring some aspect of their energy usage and efficiency in their facilities, the excuse that it's just too difficult will soon enough lose all credibility.