Cloud Based Computing and the Breathing of the Biosphere

Recently, Microsoft Research showcased an environmental project that offers researchers data resources for detailed climate science study, without having to spend money on computer hardware to process the detailed data that are necessary to examine the science. Microsoft was talking about the Azure Cloud system.

At the heart of it, the system combines a state-of-the-art biophysical modeling approach with a detailed cloud-based dataset of satellite imagery and ground-based sensor data. It supports a carbon-climate based scientific synthesis and analysis that reaches a global scale. This approach enables scientists from different disciplines to share data in order to help them understand how ecosystems behave as climate change occurs.

Studying the environment is very difficult, and as Dennis Baldocchi, professor of biometeorology at the University of California, Berkeley said, “In trying to study the environment we’re essentially trying to paint a picture by numbers, and that picture is a movie. Every day and every second it flips to another scene.”

This, in turn, leads to the problem of data processing. In such a case, the promise of factual processing and data analysis can be very expensive. It could tie up hundreds of computers for long periods of time; it may even necessitate the purchase of more equipment to make the process happen.

Here is what has to happen: The collection of data occurs from satellite imagery that includes more than 500 FLUXNET towers. This forms a global network of field-based sensor arrays that measure fluctuations (flux) of carbon dioxide and water vapor. However, picking up the data is one thing; next the data must be analyzed in a fine detail; even so far as down to a single kilometer, or as high as a global scale. In addition, data can also be scaled by time. This could be an immediate picture measured covering a satellite’s five-minute sweep of a defined area or to the complex changes tracked over a long period of time, years, even a decade or more.

So how is the data analyzed?

The computational system comes from MODISAzure. This is a downloading stream that also does processing and reduces diverse satellite imagery to a workable level. The Windows Azure platform will deliver the results of the extensive cloud computational power to the desktops of researchers. Furthermore, it will not require the investment or maintenance of a large and powerful supercomputing system. For example, the Berkeley researchers have so far used over 250 CPU hours of processing time to date, which is a small amount by most standards of supercomputing; this, however, is a significant amount of time for individual researchers with only laptop or desktop systems. 250 CPU hours would translate into 10 solid days at 24 hours a day of solid data analysis. But, the MODISAzure cloud takes that problem non existent. The Cloud processes the data, and sends the results to the scientists for scientific analysis.

If this is one example of how the Azure Cloud can help scientists, it could also help economists, finance specialists, medical researchers, astronomers, and other scientists.

In fact, with a cloud-based system, researchers are able to focus on scientific questions without having to worry about the cost or logistical issues of managing computers. The focus will be science, not IT. This means that the cost of computing power fades, and collaboration with researchers will be more important. This may lead researchers to use other Microsoft tools like Sharepoint, which is a collaboration tool. For science managers that makes the most important resources to observe and keep track of are the time and expertise of the people doing the research. Sharepoint can do that as well. It can help in the writing of articles making revisions and keeping track different versions. It can restrict access to users, and allow other to make modifications. There are many possibilities but it starts with the cloud.