Bill Weihl, Google’s green energy czar, says the computer you’re likely using now is a bit like a toaster. It takes in energy and produces heat. A typical network server is more energy-efficient, but Weihl says it’s usually loaded with cheap parts and still wastes a third of the power it consumes, which emanates as heat. That’s not really hot enough to toast bread, and probably not enough to bring a teapot to boil. But a key issue for the industry is that it takes as much electricity to cool a typical data center as it takes to run the servers inside it.

So it was no minor accomplishment for Google to recently cut its energy use roughly in half, which is what Weihl told a conference in San Diego yesterday afternoon. You could even say it makes Google a green giant in the Valley (ho, ho, ho). It’s also worth noting that Weihl says engineers at the Mountain View, CA, technology colossus achieved the reduction by rejecting industry “best practices” and designing their own servers and data centers.

The two-day symposium on “Greening the Internet Economy” in itself represents a growing recognition by the ICT industry (Information and Communications Technology) that the computer and everything it’s connected to consumes enormous amounts of power. Also noteworthy is the fact that the California Public Utilities Commission co-sponsored the conference with U.C. San Diego, because of the role the agency has played in recent years in promoting renewable energy and curbing global warming. Weihl, a former computer science professor at the Massachusetts Institute of Technology and former CTO of Akamai Technologies, got top billing as keynote speaker on the opening day.

Weihl says Google has committed itself to being carbon neutral by reducing power use, relying on renewable energy and investing in projects that cut greenhouse gas emissions. It’s seen by many as a noble goal, but going green also impacts Google’s bottom line. For the past seven years, Google has been designing its own servers because it saves money over the long run. Weihl says the company spends between $20 and $100 more for each server it buys, depending on size. But he says the extra cost pays for itself in reduced energy costs within six to 12 months.

The Internet search engine giant also designs and builds its own data centers, using evaporative cooling to keep its systems running. Citing figures from the Environmental Protection Agency, Weihl says the typical enterprise data center has a power usage effectiveness, or PUE, of 2. That means a facility consumes twice as much power as the computer equipment inside it. Weihl says Google has reduced its PUE to 1.2.

“Overall, through the efficiency work that we’ve done by … Next Page »

Comments (3) | Reprints | Share:          

Apparently, a lot of executives have a mystical faith in innovation, believing that it will pay off somehow, even if they can’t see the results from previous investments.

That’s one conclusion from an analysis published today by local management-consulting leader Boston Consulting Group. Only 46 percent of executives BCG surveyed this year say they’re happy with the return on their innovation investments—yet a full two-thirds plan to increase spending on innovation.

But crossing one’s fingers probably isn’t a viable long-term strategy. The BCG report argues that companies should be more careful about how they measure innovation—tracking not just how much they invest, but also quantities such as the percentage of new ideas funded, the number of projects killed at each milestone, and (duh!) profitability.

“Poor measurement practices translate into bad or incomplete information, wasted spending and, ultimately, a lower return on the investment in innovation,” said James P. Andrew, the BCG senior partner who ran the study, in a statement accompanying the report. “And when the majority of companies are already less than satisfied with that return, lack of measurement makes things much worse.”

The report found that 60 percent of companies use five or fewer metrics to track their return on innovation, “well short of the number necessary,” in Andrew’s words. And only 28 percent of executives said their companies use these metrics to parcel out employee incentives and rewards.

The solution? Andrews recommends that companies use a “cash curve” model that tracks the cost of an innovation project over its entire lifetime, from startup (when expenses are largely determined by the time it takes to get a new product to market) through launch (when the time required to scale up to large volumes becomes the key factor) and postlaunch (a period marked by ongoing support costs).

Not coincidentally, the cash curve model is the subject of Andrew’s recent book Payback: Reaping the Rewards of Innovation (Harvard Business School Press, January 2007). Booklist, the book review publication of the American Library Association, called Payback an “infomercial” for BCG, but also said it offered “a disciplined and analytical approach to determining how much and where to invest” in innovation.

Comments | Reprints | Share: