Putting Your Computer Where Your Power Is
As the popularity of green computing attests, energy and climate security are major concerns. Energy prices are going up and fossil fuels are problematic. These finite carbon-based resources may last another couple hundred years or more with advanced capture techniques, but the specter of climate change demands swifter action. Given these realities, the status quo will no longer be sufficient. New approaches are necessary.
Thanks to the Internet and decades of software and middleware advances, computing can be delivered much like a utility, where clients pay only for what they use. In this analogy, the Internet is likened to the power grid. In fact, that's where grid computing got its name.
New Mexico-based startup SELC Corp. has taken the utility computing concept and turned it on its head by asking: what if instead of sending power to the computer, it were possible to send the computer to the power? The company asserts that not only is this setup possible, it is preferable, and will enable truly green computing.
SELC (pronounced SELL-SEE) stands for Smart Energy Load Centers. Founded in October 2012, this innovative startup is the brainchild of Dr. James Stalker. In late 2009, Dr. Stalker was inspired by his attendance at an energy conference in Cape Town, South Africa, to come up with a new concept for renewable energy development. A provisional plan soon followed and this became the template for a full patent application, submitted in April 2010.
Still pending, the patent lays out a new method of powering computers – one that is radically different from the current transmission-based model. The concept starts with Smart Energy Load Centers (SELC), the energy optimization system that Dr. Stalker developed. Stalker asserts that locating the most power-intensive parts of the computing system (personal computers, cloud/server clusters, etc.) next to the energy source will result in a dramatic energy and cost savings, which can be passed onto the end user.
The project is still in the very early stages, but the CEO's goal is to complete a working site in 27 months.
In an interview with Green Computing Report, Dr. Stalker described three main challenges to renewable energy, which SELC aims to solve.
The first is the power grid requirement. Utility-scale projects need the power grid nearby, and even when it's nearby sometimes there is no capacity, says Stalker. The second challenge is obtaining a long-term power purchase agreement (PPA), which is usually only an option for the largest organizations, those with high-volume purchasing power. Lastly, renewable energy is chaotic, and stabilizing the peaks and valleys (via some storage mechanism) is costly. The cost of stabilization can be as much as the energy production cost itself, which in the case of renewables is already higher compared to conventional energy production.
Dr. Stalker outlines the ways that SELC's grand plan tackles these three challenges. First, it removes the power grid problem by consuming most electricity (90-95 percent) locally (i.e., at the generation site). Second, because SELC will become the local utility for the developers/investors, long-term PPAs won't be necessary. Third, the computing systems will be modified to create cost efficiencies that will offset the stabilization (energy storage) costs. The main modification will be the elimination of redundant parts, such as the power supply units. All computer units will be provided with a constant DC power source; they won't be connected directly to the energy site (e.g., the wind or solar farm).
Dr. Stalker describes the initial user base as "small, medium or large organizations, not necessarily datacenters," with somewhere between 200 to 20,000 personal computers. SELC will replace these full-fledged PCs with a thin client that uses about 10 percent of the previous energy load. The remaining 90 percent of the energy used in computing will be consumed at the energy site by SELC's modified computers.
This energy split is a key claim of SELC's patent. Dr. Stalker says that with today's technology an energy split of 65/35 is possible, but he is confident that R&D will achieve a 90/10 split in two years. SELC's eventual target goal of >95/<5 extends beyond that two-year R&D period.
Exactly how the computing will be delivered has yet to be determined as SELC plans to outsource connectivity-related tasks. The remote computing technology certainly exists. What's more, the bulk of business applications are not particularly IO-constrained and are therefore good candidates for this kind of network-based computing model.
When the discussion veered toward computing specifics, the SELC CEO emphasized that the patent is on the energy side, and while he has identified computing as a primary use case, he says the concept makes sense for a range of verticals, such as 3D printing and other manufacturing processes. Eliminating the grid will eliminate a lot of the cost, notes Dr. Stalker.
In fact, SELC Corp plans to deliver clean, green stable energy at a lower cost than the utility rates. Utility rates are only going up, and SELC believes it can offer a fixed rate for the life of the contract because its energy source (a combination of wind and solar) is effectively free. The company is quoting rates of 10.3 cents per kilowatt-hour for 100 percent renewable, while the traditional retail rates are 10-25 cents per kilowatt-hour depending on the market.
Given the recent studies about the energy efficiencies of a cloud model, perhaps the ultimate green IT solution will be a combination of large clouds located next to renewable energy sources. Cloud providers like GreenQloud and Facebook are creating green cloud ecosystems by siting their clouds next to renewable energy (or putting renewable energy sites next to their clouds).
SELC is taking a different approach by coming at it from the energy side, as the CEO explains:
"We have some claims on the computing side, but I don't want people to say 'this looks like cloud services.' I tell people there's a much broader claim. It's the energy side: eliminating the grid, eliminating the utility, and stabilizing the power source without doubling the cost of energy production. Those are the things we have claims for in our patent-pending technologies."
As SELC is a New Mexican company, the first projects will likely be in New Mexico and will use a hybrid of wind and solar – 80 percent wind and 20 percent solar. Dr. Stalker expects the 200MW inaugural site will cost about a half-billion-dollars to build. With over two years to go until their official launch, SELC plans to use the time to develop R&D and line up partners and investors.