Recently I saw a commercial for a popular mobile services company on TV that claimed we-the-consumer’s “right” to unlimited data. Such advertisement sloganeering is testament to how, in our interconnected society, data serving has become a common utility, an expected right and a shared organizing aspect of modern life. The energy intensive nature of data centers combined with ever increasing demand for always-on servers pushing “unlimited data” to the growing information consuming masses makes datacenter power and cooling monitoring increasingly crucial for efficient and economical facility management. I recently spoke with Lisa Mandell, president of Trendpoint Systems Inc., a manufacturer of utility–grade power and cooling monitoring systems. Trendpoint exists on the fine dry edge of datacenter management, with a product line that includes the EnviroCube cooling management device, EnerSure branch circuit power meter and the EnureSure Bus busway power meter. Power metering and cooling management go to the heart of maximizing datacenter efficiency and capacity. “When you’re trying to manage your infrastructure,” says Mandell, “you have to get information coming from somewhere, and obviously it comes from meters.” You can’t manage what you don’t measure and the granularity and flexibility of Trendpoint’s innovative product line puts the company in a singular position to bring greater efficiency in the face of an exploding industry, both in new builds and retrofits. Trendpint’s target market is co-location companies like Raging Wire, government agencies (the FAA is a client) and large commercial accounts such as VMWare and Loral Space Systems, and several others who Mandell says prefer to remain anonymous. “It’s not the kind of B–to–B that lends itself to advertising,” but even so, Mandell is optimistic for future growth and confident that the market for Trendpoint is “wide open.”
Efficiently managing the resources of a datacenter can broadly be defined by three primary concerns:
- Minimizing the cost of power: One-quarter to one-third of energy in a datacenter is wasted through inefficient cooling. “Through monitoring you can map how much energy you are spending cooling versus how much heat you’re actually generating by running all the servers,” says Mandell. “So you can, in theory, determine if you’ve got an air-conditioning unit that’s pumping out cold air all day but not cooling anything–not tied to anything that’s generating heat, and you can switch those off and save energy …it’s about matching cooling needs to the heat being generated and it’s really that simple.”
- Managing to capacity: “So you’ve got all these server racks,” says Mandell, “and the server racks are all full. When you’re monitoring the energy – the power going into a rack – you know how much excess power you have to put more servers in the rack. Maximizing your capacity is first and foremost in the data center.”The increase in the power density of data centers makes it important for operators to maximize the “amount of equipment you can shove into the space and provide power to them. Mandell explains that data centers originally built with a target capacity of 35 watts per square foot are now building out to upwards of 2000 watts per square foot. This increase in power density makes it important for operators to maximize the “amount of equipment that can shove into this space and provide power to them,” says Mandell. “So this lends itself to doing capacity management, as well as saving power by understanding where your power is going, and lastly it enables clients –especially co-location companies – to bill their customers for specific consumption.”
- Uptime: “… keeping all the equipment up and running [is truly first and foremost]. And power is one of the reasons why something would shut down, if you’re overloading a circuit. So if you’re managing to the amount of amps that are actually being supplied to the servers you can see and set alarms” providing an opportunity to deal proactively with a power anomaly before something goes down.
Vendor agnostic flexibility
Almost all datacenter facilities have a mix of busway systems supplying power to the server racks. Trendpoint’s busway metering systems are “vendor agnostic” and will work with any busway system on the market. Illustrating the point Mandell gives the example of a recently acquired major multimillion dollar client: “It’s a customer that had numerous datacenters, they’re all busway and they use Starline in half of them and they use PDI in the other half, and they want to standardize on one metering platform. So even though PDI and Starline have their own meters, obviously PDI’s meters to fit on Starline and vice versa.” And the situation is not unique to this one client Mandell explains:
“It’s hard to imagine any customer having only one type or only one vendor for power distribution, whether it be busway or traditional, and that gives us a huge advantage. We can hook into any DCIM (datacenter information management) system, they don’t have to take anything down to install us and we have all the flexibility – within each different power panel – to have any size circuit that they have within one power panel. And we are the only one’s that do all of that.”
Trendpoint’s product line likely seems a bit esoteric and geeky to the non-IT oriented among us (it took a couple of conversations with Mandell for me to really begin to wrap my head around some of this), but not so for facility managers working to bring costs down, conserve energy and maximize capacity. And the devil is in the details, the nuts and bolts of how to manage for efficiency and capacity in an inherently power-intensive industry that will only continue on its path of rapid growth in an era where “unlimited data” is as ubiquitous and expected-on-demand as the electricity that gives it life.