US scientists get free cloud on-ramp
Microsoft and NSF plugging Azure
Microsoft and the US National Science Foundation have announced an agreement that will provide free access to cloud computing resources for select NSF-funded researchers for the next three years.
This was discussed by Microsoft's corporate VP for tech strategy and policy, Dan Reed, in a blog. Instead of buying supercomputers and massed ranks of storage arrays, the lucky scientists get to use remote Microsoft Azure data centres full of Windows/Dell servers and storage so that they can run compute-intensive algorithms on masses of data.
Reed suggests the data mining of weather sensor-gathered data as an example application.
Microsoft is going to improve its Azure cloud computing access and interface tools by having its researchers and developers "work with the NSF grant recipients to equip them with a set of tools to assist them in expanding their research into the cloud".
A big advantage claimed for cloud-based computing by scientists is much easier sharing of data. Microsoft and the NSF provided a canned quote about this from Dave Patterson, Pardee Professor of Computer Science at UCAL, Berkeley:
If you want to collaborate with people, especially people across distances, it makes a lot of sense to keep your data in the cloud rather than in your own ivory tower. The next step is the sharing of programs to manipulate the data in the cloud and the sharing of computer simulations. Sharing is trivial once a researcher gets it working in the cloud compared to shipping files and then to try to get data and programs installed in and working in other places. The way cloud computing works, if I’m getting my work done, there’s no reason everybody else can’t use it as well.
Eligible individual researchers and research groups will be selected through the NSF’s merit review process, and the relevant part of the NSF to which researchers should apply for funding is the Directorate for Computer and Information Science and Engineering (CISE). ®
Sponsored: Hyper-scale data management