Monday, August 20, 2007

IBM forms 'Big Green' think-tank

Helping the environment and business profit are not mutually exclusive goals, and through its new think-tank, Big Green, IBM is pioneering new technologies and new management models to combine the two.

Peter Williams, chief technology officer for Big Green, explained how IBM has been a company with roots in environmental awareness for well over 35 years. Years ago, IBM came up with manufacturing technology that reduced the amount of perflurocarbons (PFCs)--a greenhouse gas much more damaging than carbon dioxide--and gave that technology away to the industry.

"Big Green came out of strategy 2.0, the World Innovation Jam. We took a question, 'What are the innovations that matter to our company?' and we threw that question out to every IBMer, their families and their clients," he explained.

Over 140,000 "IBMers", as they call them, participated in the innovation jam and 40,000 ideas for Big Green were put forward. These were later moderated down to just 10.

One project was to leverage IBM's expertise in network integration and water management. Another was to use IBM's semiconductor expertise to make better, more efficient photovoltaic cells and desalination membranes. Other projects include better energy transmission, carbon management, all in all a delightful mix of pure technology and ICT at work for the betterment of the world we live in.

A river in New Zealand is the first test bed for IBM's integrated water management solution. Big Green deployed 259 sensors along a short river to monitor, learn and predict water flow. This information is processed along with weather forecasting and then fed to farmers. The weather forecast by IBM's super computers is very fine-grained, down to a one square kilometre grid. Further information on soil and crops is added to the model and what comes out of this system is a recommendation for the farmer of what to plant and when.

While information on the pilot project was made available online, it could just have easily been printed and handed out to farmers in rural Thailand.

Further work is being done along the Mississippi, Amazon and Yangtze rivers.

Reducing the carbon footprint for data centres is also a major issue. But how is reducing carbon different from reducing energy? For Big Green, the innovation is in the models and in providing tools to do what-if analysis with business decisions when it comes to carbon dioxide emissions.

"If we can align green with the economics motive, then everyone wins. There are many things that we can do to reduce carbons, but there are other things like improving customer service that might or might not reduce carbons. We need to help business manage these trade-offs and provide information to those who make decisions so they consider carbon," he said.

IBM has for a long time been looking at the carbon impact of the packaging of each computer it sells. All things being equal, less packaging means more can be put into a freight container and the carbon footprint drops. But if the packaging is reduced too much, then there will be breakage and the carbon footprint of repairs needs to be added.

So while carbon cutting is today very closely aligned with business as usual, some aspects are causing people to re-think their business processes. For instance, a modern automotive factory using just-in-time manufacturing processes is a very carbon-intensive process. By adding buffer stock, the number of trucks, buses and trains making deliveries could be lessened. But that buffer stock has to be illuminated, heated and that in turn increases carbon emissions.

In the power arena, Big Green has two projects. One is to use IBM's advanced semiconductor technology to create better, more efficient solar cells.

"It's not about using a 45-nanometre process or the latest 300 millimetre fab. We can do it with normal silicon; we can improve the layers, the way connectors are made. We have a silicon process that will lower the cost of manufacturing," he explained. "The ultimate expression is the multi-junction cell. By using multiple layers of different materials that are sensitive to different wavelengths, we can allow the cell to get much higher efficiency."

Today's best solar cells return around 18 to 19 percent efficiency. Williams says that the silicon in IBM's labs can produce significantly more power and at the same time be much cheaper to produce. How much more efficient and how much cheaper? Williams said he is not allowed to give figures just yet, but his enthusiasm suggests we may be in for a pleasant surprise.

Big Green is essentially an incubator of ideas which will be later handed over to the various IBM business units to develop and sell. The core Big Green team only has five people on it, but more than 25 other IBMers have been given permission to help with half of their working time on the project.