Human's face quite an enormous challenge of studying and understanding the universe. Global efforts in this direction have been going on for several decades. Supported by reliable technology, astronomy, therefore, has come a long way. Radio telescopes are the mainstream modes for studying the cosmos. The International Centre for Radio Astronomy Research (ICRAR), based in Perth, Australia, is working on developing the biggest radio telescope ever built, known as the Square Kilometre Array (SKA).
In a single day, SKA telescope is expected to capture and process as much data as currently produced by the digital-world in a year. This data will be used to map the sky for studying the universe. Therefore, ICRAR formed a community-computing-initiative called theSkyNet to address the challenge of computing the mind-boggling amounts of data produced as a result of sky-imaging. Crowd-sourced computing experiments usually face fatal operational problems. Hence, ICRAR needed to perform experiments to process sky-images produced by currently operational Pan-STARRS1 telescope as a part of theSkyNet project. Running these experiments in a flexible and cost-effective way was the biggest challenge. Because, the valuable results obtained from these experiments will help in successful launch and functioning of SKA.Efficiently managing the data from theSkyNet community was also one of the major challenges.
The on-demand and scalable nature of Amazon Web Services (AWS) made it possible to provide the right resources needed by ICRAR to analyse astounding amounts of sky-imaging data. These data-science-based resources help in processing, storing and disseminating data churned out by theSkyNet experiments. AWS have, hence, eliminated the need of a dedicated supercomputer for running these experiments. AWS have provided a cloud-compute instance as a data-processor and a network file server. ICRAR further started storing and archiving the necessary data using AWS storage solutions.
AWS enabled ICRAR set up theSkyNet project in just 4 days. Consequently, ICRAR saved a lot of their valuable time.
A monthly processing of 150 GB of sky-imaging data and 400 GB of storage is now available to ICRAR. Hence, this helps in simultaneously analysing data from 400-500 galaxies.
ICRAR does not need any new servers to address data-overload problems on theSkyNet infrastructure.
To read more about the interesting journey of ICRAR’s migration to AWS, please visit http://amzn.to/1HRVOM1