Internaut | Floating data centers: Good for government?
- By Shawn McCarthy
- Oct 30, 2008
As the federal government seeks to make its next leap forward in affordable computing, perhaps it needs to borrow a few ideas from the realm of science fiction. That's what Google seems willing to do, and it will be pretty exciting to watch its new ideas unfold.
If you're a fan of cyberpunk novelist William Gibson, you probably already know that moving data facilities offshore, literally, is looming somewhere in our future. But that target has always seemed at least 50 years down the road.
Well, dust off those mirror shades. Google is accelerating the idea by applying for a patent to develop a "water-based data center." If built, it would include a floating data facility, an onboard set of generators powered by wave action, plus a cooling system that uses sea water and wind power. The company's early drawings also highlight separate ideas for shore-based facilities that can be powered by daily tidal fluctuations.
It's probably fair to address the wave power generation idea separately from the floating data center idea. Both have their advantage, and either could find a niche separately from the other. Yet together, they hold a vast amount of potential.
So what's the government connection? Well, let's start with the proposed power generator that Google may use. It's produced by a company called Pelamis Wave Power Ltd. The company's first full-scale wave farm was built in Portugal this year by a Scottish company called Ocean Power Delivery. While the installation eventually will serve commercial power needs, the project itself was made possible by support of the Portuguese government.
Yes, there are still many engineering questions that still have to be answered. But the federal government should be supportive of this effort because Google's idea has the potential to solve several problems that the government itself faces when it comes to building and maintaining large data centers.
1) Data center power needs to be cheap and reliable. But today, with rising energy costs and a vulnerable power grid, no electrical supply can be taken for granted. Yes, temporary generators are an alternative at any site. But a floating data center with wave power takes self-sustainability to a new height'or depth, as it were.
2) Expanding an existing data center can be complicated. If extra land is needed, it's costly. Dealing with local property owners and zoning laws can trigger unexpected legal bills. Plus, is local Internet bandwidth cheap and reliable enough, now and for the long term? While a floating data facility also would be very expensive, costs should eventually be contained and predictable. And you always have the option of moving a water-bourn data center to where the best bandwidth is available near the shore.
3) Traditional data centers are vulnerable to natural disasters. This is true for floating data centers, too. But a floating center could be temporarily switched off and relocated, say, out of the path of a hurricane.
4) Modern databases, applications, servers, etc. need to have built-in redundancies. But space is always an issue, and geographic separation is often required. Now think of that issue with floating centers. Need a redundant data center? No problem. Just park another floating facility a few miles (or a few hundred miles) away. Instant redundancy!
5) Building out any data center is a crapshoot because of construction delays, costs that vary by region, and unique design issues. But if the floating data center idea ever takes off, it should eventually provide quick predictable deployment and a known cost structure. Hopefully it will even come with a quantifiable return on investment.
6) Upgrading servers, hubs and system management units at a data center can be a logistical challenge. Google has promised to make its floating computers available in groups that are integrated into modules that can be swapped out by crane as needed.
The bad news: Ask any Navy IT manager and he will tell you, a marine environment can be a harsh one. Humidity is high, corrosion is a constant problem, the relentless motion of the ocean can take its toll on equipment and wires, and fighting fires on a ship is a serious challenge. Yet, the multiple high-tech systems on today's Navy ships prove that the problems are not insurmountable.
Google's floating platforms, if they become a reality, offer one additional value for the government. It will help bring the idea of electricity generation from wave power and tidal power from early adolescence into maturity. And many other forms of electrical generation are worth exploring right now.
So if Google's idea works, what's next, an island of plastic garbage that declares itself to be a new country, complete with its own currency and a top-level domain name? Fun to think about, isn't it?
Details of the Google's idea can be found by visiting the U.S. Patent and Trademark office Web site and searching for Patent Application No. 20080209234. (Be sure to search for patent application, not issued patent.)
Details on the Pelamis Wave energy Converter can be found here: http://www.pelamiswave.com/.
Shawn McCarthy, a former writer for GCN, is senior analyst and program manager for government IT opportunities at IDC.