ENERGY STAR for UPS

Speaking of ENERGY STAR, the EPA released a framework document for the newly announced Uninterruptable Power Supplies (UPS) specification last week.  UPS, like computer power supplies before them, lack industry standard measurement procedures to specify their efficiency.  As the market for energy efficient data center equipment grows UPS makers seem to be increasingly marketing the efficiency of their devices, but manufacturers usually specify 100% load – a condition that a UPS will never actually operate in because many UPS are critically underloaded.  Also similar to server power supplies, many UPS are operated in redundant configurations where the load is split between two UPS in the case that one fails.  This means that a UPS in this configuration could only hit 50% load, max.  The efficiency of power conversion equipment tends to fall off below 50% load, so it’s important to measure and specify the efficiency of loads below 50%, because this is where a lot of this equipment is actually running.

To illustrate the point, here’s a chart of power supply efficiency curves from when I was working on the server specification, which I stole from the ENERGY STAR website:

For servers, EPA specified efficiency all the way down to the 10% load condition because available data indicated that that’s where a lot of the redundant power supplies were being operated.  My guess is that ENERGY STAR will be doing a similar thing with UPS, and then the industry will have a way to compare the efficiency of different UPS solutions across much of their operating range. This should be a great help to utilities looking to get verifiable savings through offering incentives or rebates for more efficient UPS.

EPA is also continuing the trend of pushing for standardized reporting requirements (through a power and performance data sheet) and for real-time power and temperature reporting over a standard network.  This is also similar to the V1.0 server specification and what is being proposed for data center storage equipment. EPA is looking to add similar requirements for all data center equipment so that data centers can be operated more efficiently when the managers have better access to data on what’s actually happening in their data center. The power and performance data sheet will also be helpful for proving the specifications of equipment when applying for rebates and incentives.

Interested stakeholders can download the new documents here, and offer comments by April 2, 2010.

Utility and ENERGY STAR Collaboration for Improved Specifications and Programs

I spent Tuesday reacquainting myself with my old friends over at the ENERGY STAR program by attending the ENERGY STAR information sessions for Servers and Storage that preceded the Green Grid Technical Forum. It was interesting seeing things from the “other side of the podium” by being a stakeholder at these meetings instead of being in my old role of assisting the EPA on the development of the specifications.

Status of the Specifications

In terms of status, there seems to be some significant work to be done on both specifications, but as usual EPA is asking the right questions. For both specifications the question is how can you quantify the generalized “efficiency” of the product, or the amount of useful work and performance you get from a system for a given energy consumption? This is the ideal outcome of this process – what everyone wants. As Andrew Fanara (the lead representative of the EPA) said, “I’d also like to ride a unicorn to work”. Meaning that it would be impossible to get a perfect metric, so for now we need a method to rank IT equipment by it’s efficiency, but don’t expect it to be perfect. There’s hope that we’ll get there eventually, but it will be a long processes, as there are a lot of details to be worked out.  The server specification feels like it’s getting closer – they’re currently working on version 2.0 so they’ve been asking these questions for longer – but there’s still a lot of work to be done.  One good thing is the EPA is showing that they’re willing to think a little differently about these products.  I think this is necessary because the complexity of these products and the subtleties of this market make theses specification development efforts very different from many other products the EPA is used to dealing with.

Utilities and ENERGY STAR

I’ve been feeling that there is a gap in thinking between the EPA and the utility industry, and the funny thing is that I think they really need each other. The utilities are constantly looking for new savings opportunities and it’s a lot easier for them to develop effective programs if they are built on the back of good efficiency specifications.  What the EPA needs are stakeholders with a voice to help drive these specifications towards increased levels of rigor for energy savings.

In addition, there needs to be a closing of the gap between the needs of utilities and the output of the ENERGY STAR program. ENERGY STAR should be producing specifications that can easily be adopted for utility programs.  This should be a high priority for ENERGY STAR, but it feels like the current process is to produce the specification without utilities in mind, and then try and adapt the result to a program.  If utilities want to play in this space, they need to be at the table learning about this industry and helping drive the agenda.

Right now vendors dominate the ENERGY STAR meetings. The vendors are extremely knowledgeable, but obviously biased towards their own products and agendas. The meetings often result in vendors standing up and talking about what isn’t possible or what EPA shouldn’t do. What the efficiency community needs are stakeholders at the table telling EPA what they need to help make these specifications useful tools to leverage for energy savings. The way to speed up this process and to keep ENERGY STAR specifications relevant is to have efficiency advocates help drive the process.  This may involve helping generate data and providing some technical resources. This will be expensive, but if the utilities (and other EE advocates) pool their efforts this should be cost effective and will help ensure a useful product for adoption.  The more utilities bring to the table, the more influence they will have.

The thing is that EVERYONE should benefit from useful ENERGY STAR specifications and effective utility programs that leverage these specifications:  ENERGY STAR can further increase their growing relevance in this emerging market; utilities can run influential and cost effective programs to meet their goals; and vendors can market more efficient product offerings.  It’s a win, win, win.  We can no longer let the voices of manufacturers, who seem afraid of being left out of the party because of inefficient product offerings, dominate this conversation. It’s time for utilities and other advocates to team together and help influence this process to get a leg up in this market.

Data-Driven Prescriptive Incentives for Data Centers

I’ve spent the last few weeks reviewing information on utility data center energy efficiency programs for a presentation I’m putting together for AFCOM Data Center World, and have been struck by how few prescriptive programs (or “deemed measures”) are available for data center equipment. The few programs that do exist seem to vary widely and are distributed among different utilities around the country.

The most important information you need for creating these programs are data on which to base your assumptions and calculations, including, data on the typical products in the market (or the baseline) and data on the more efficient offerings.  The delta between these two establishes the energy savings on which to base an incentive.  This was similar in my previous work developing specifications for the ENERGY STAR program for the EPA – the biggest problem was access to quality data of sufficient quantity to really understand the energy use of the products. This is a great strength of the ENERGY STAR program, because as a trusted third party they are able to pull in data from a number of different current, and sometimes future, products to get a real sense of how energy is used across different manufacturers.  When they are pulled together these EPA data sets often seem to be some of the best publically available data sets on the energy consumption of these products – a resource that is useful for the program, as well as for other advocates outside the program.

It seems that the utility industry lacks a similar mechanism to collect sufficient data to develop these prescriptive incentive programs for data center equipment.  The primary source of data seems to be data collected through demonstrations and custom incentive applications.  But the utility industry needs a large amount of data to maintain confidence that the prescriptive programs will deliver actual energy savings that they can reliably claim for their programs. However, the utility industry, like ENERGY STAR itself, has only recently taken the plunge into the data center industry.  A lot of programs have a random assortment of incentives they’ve given out for data centers – an efficient UPS here, an economizer there, a few virtual servers in the mix – and do not seem to be reaching the critical mass needed to gather the quantity of information needed to effectively develop prescriptive programs.

I think this turns into a chicken and egg problem. Utilities sometimes find it hard to get traction on their data center programs because they do not have the prescriptive programs that make it easy for the customer to participate, but without the data from participation in the programs they do not have the information needed to develop the prescriptive programs.  As usual, available data seems to be a bottleneck.  What is needed are some central depositories of data with mechanisms to develop intelligent incentives based on that data.  A lot of different groups and organizations have the potential to work toward this goal (and I believe are doing so), and it is an important goal as more prescriptive incentives would certainly help capture some of the energy savings potential which we all know exists in this industry. I’d be really interested to hear about any potential efforts in this area, so if anyone knows about anything fill me in!