Data centers make an interesting, if challenging, target market for integration into the smart grid. Sure, they’re a huge source of power, running about the smartest power loads in existence -- lots and lots of servers, routers, switches, and other IT equipment. That should make them a natural fit for the network of IT that’s connecting utilities to their customers to better manage power across the grid.

Data centers also tend to have a pretty hands-off attitude when it comes to hooking up to the local utility. Most data centers have installed expensive batteries and backup generators to take over during outages -- but they’re not very interested in turning over those resources during peak power times or other such grid emergencies, since that’s when they’re also most likely to be facing their own power crunch.

But rising power prices, combined with the IT world’s insatiable hunger for more server space, are bound to push data centers and utilities closer together. Beyond lowering power bills, data centers face key pinch-points in capacity that need to be handled right away -- some customers may need to increase computing capacity in the same crowded space, while others may need to increase capacity without exceeding the amount of power their local substation can provide.

Luckily for the data center world, it appears that their IT needs and their energy demands can align nicely for savings for themselves and the grid. According to an August report from Lawrence Berkeley National Laboratory (PDF), data centers can shave power bills by up to a quarter, simply by managing their IT resources more intelligently -- and with the grid’s needs in mind.

The study included funding from the California Energy Commission’s Public Interest Energy Research (PIER) Program and utilities Pacific Gas and Electric and San Diego Gas and Electric. It also saw participation by two data center efficiency startups, Santa Clara, Calif.-based PowerAssure and Sacramento, Calif.-based SynapSense, which are both working with data center partners in the state.

According to the study, “With minimal or no impact to data center operations, a demand savings of 25% at the data center level or 10% to 12% at the whole building level can be achieved with strategies for cooling and IT equipment, and load migration.”

The report came with a long list of caveats, however, including the fact that it only tested four data centers. It also noted that not all of the responses used for maximum energy savings would be appropriate for “mission-critical” data centers like, say, the ones that control Wall Street or the Pentagon.

It also stressed that “load migration,” or moving computing loads back and forth between data centers to maximize energy market savings and revenue opportunities, would require an underlying level of IT sophistication that not all data centers have.

“If the applications are data-center independent -- if they can switch from one to another seamlessly, which they have to do for reliability anyway -- then you have enough flexibility to play in the energy market,” Clemens Pfeiffer, CTO of PowerAssure, said in an interview last week. “On the other hand, “If the application doesn’t support it, you can’t do it.”

In other words, data centers are already making the IT infrastructure improvements they need to play into energy and grid markets today -- if only they have the tools to recognize it. Becoming aware, in turn, can lead to grid-facing opportunities, both to reduce cost and to generate revenue.

So how have data centers taken up the challenge? We’ve seen a number of projects building the links to allow data centers to interact with the smart grid at large, such as Cisco’s project with NetApp and Pacific Gas & Electric in 2008. That project paid itself off in utility rebates as well as power bill savings.

But the next level of data center energy optimization -- the kind that brings load migration and other more advanced concepts into play -- is far less common. After all, this technology is all brand new, and most data center operators are leery of turning over control to any system that could negatively affect uptime.

Even so, we’re seeing more and more integration of energy data into the way data centers are run today. IT giants like HP, IBM, Intel, Cisco, Microsoft and Oracle are improving server performance per watt, and integrating energy data more closely into their data center infrastructure management (DCIM) platforms. Big grid players like ABB, Siemens, GE and Schneider Electric are also making forays into the data center. Then we’ve got startups like PowerAssure, Vigilent (formerly Federspiel Controls), Sentilla, Vigilent and JouleX, to name a few, each bringing their own combination of capabilities to market via partnerships in the industry.

PowerAssure, for its part, has been working on data center-smart grid integration with Virginia’s Dominion Power since late last year. Earlier this month, it announced a partnership with iTRACS, a maker of data center mapping and visualization software, to provide its suite of software as a cloud-hosted service, or as a turnkey platform for customers concerned about keeping data center data under strict security, Pfeiffer said.

Other startups are linking up with companies that are designing the next generation of high-efficiency gear for the data center. JouleX, the Atlanta-based startup that works with customers including Cisco, Intel and VMware, has also worked closely with Calxeda, maker of ARM-based high-efficiency servers, for example. It has also worked with Cisco on a powered-over-Ethernet (POE) LED lighting system, similar to that deployed by startup Redwood Systems in Facebook’s super-green data center in Oregon.

Underlying all these technical approaches is a common challenge, however, Pfeiffer said -- convincing the human beings in charge of making investment decisions that energy-smart data centers are worth the cost. “Our problem today is, we need to quantify the opportunity, then have the infrastructure to do it, and the tools to participate in it,” he said.