The project leader Dr. Maciej Malawski explained that computers with huge total computing power are running in clouds. They allow to store large amounts of data. "Throughout the world, thousands of servers process data from billions of sources, and software created by the world's most capable experts responds to these events, creating fascinating effects. Cloud computing is revolutionizing not just the IT industry, but also the way modern science is practiced" - Malawski said in an interview with PAP.
We benefit from cloud computing every day - clouds let our smartphones work instantaneously when we tell them to find the nearest gas station, automatically arrange a panorama from images, or remove vibrations from a video. Astronomers create panoramas from thousands of sky shots made by telescopes. They also use clouds for data processing, but on a slightly different scale than smartphones.
"In my phone, four processors can be used to perform local calculations. But it's often not enough to quickly process all the information, so we send data or computing tasks to the cloud to get more computing power in a short amount of time. This is parallel, distributed computing" - explained Dr. Malawski.
Other examples of cloud application in everyday life include automatic recognition of images by applications for smartphones and personal computers. If we have photos in the cloud, we can automatically search for those that have been taken in Krakow (Wawel in the background) or Paris (the Eiffel Tower). We can also search for faces of friends, and with a single command find all kid's photos. We could easily get by without such facilities, but in scientific applications clouds are necessary.
The vast amounts of data that astronomers need to analyse, based on hundreds of hours of observations of huge areas of the sky, can only be handled with the use of parallel calculations in the cloud.
"Large sky or large catalogue of observations can be divided into smaller parts, each to be calculated on a different processor or virtual machine in the cloud. The use of multiple processors in commercial clouds is expensive, so time-cost optimisation is required. Every project has a budget and you have to decide whether to analyse a larger area with less accuracy, or to get some results faster to publish them first. The calculation time is important even on astronomical scale" - said Dr. Malawski.
Modern technologies change and evolve rapidly. To use them better and cheaper, it is necessary to know them well. The aim of the project carried out by researchers from AGH University of Science and Technology in cooperation with U.S. partners from the University of Southern California in Los Angeles and the University of Notre Dame in Indiana, is to investigate and evaluate the latest cloud computing technologies from the point of view of scientific applications.
Researchers will try to make the most use of the latest, extremely flexible infrastructures for new, fascinating science subjects, such as earthquake simulations (more on this in PAP - Science and Scholarship in Poland - here). Researchers will test emerging computing clouds and accurately measure their behaviour. Based on these experiments, new algorithms will be developed to automatically optimise the use of clouds.
The amount of the National Science Centre grant is about 500 thousand zlotys. The project is scheduled for 3 years. The funds will be used to cover salaries for two doctors and scholarships for doctoral students and a masters student. The researcher told PAP that in computer science many students work while studying, and this scholarship allows to keep them longer at the university so that they can to pursue their passion in research projects. Some of the money will cover the cost of using commercial computing infrastructures such as Amazon or Google.
PAP - Science and Scholarship in Poland, Karolina Duszczyk
kol/ agt/ kap/