Microsoft unveils data-analysis tool for scientific work

Microsoft has unveiled details of a new data-analysis tool that will allow scientists to process large amounts of data.

Named Daytona, the tool is designed to run a range of analytics and machine-learning algorithms on the Windows Azure cloud computing platform.

Dan Reed, corporate vice-president of the Technology Policy Group at Microsoft, said the tool will provide more ways to use the cloud without being tied to one computer or needing detailed knowledge of cloud programming. He added that it will help allow ‘scientists to be scientists’.

The system kit was unveiled yesterday at the 12th annual Microsoft Research Faculty Summit. The theme of this year’s event is Future World and will explore new advances in natural user interfaces, cloud computing and machine learning.

Roger Barga, an architect in Microsoft’s Research’s extreme Computing Group, said: ‘Increasingly we have seen researchers wrestling with large data collections in our cloud-engagement program. They want to extract insight from these collections and they know which algorithms to use. But they need them to scale out the algorithm to process data volumes that they’ve never tackled before: terabytes of data, not just megabytes or gigabytes.

‘These are the people we see using Daytona to run algorithms over large-scale data collections to extract insights: find patterns, find clusters, find outliers and build training models so they can classify incoming data based on the data they’ve captured and classified to date.’

New releases of Daytona will be scheduled monthly and will be available to the scientific community for free.