Featured image: A computer generated image of the Square Kilometre Array (SKA) radio telescope dish antennas in South Africa. Credit: SKA Project Office.
What is dark matter? What did the universe look like when the first galaxies formed? Is there other life out there? These are just some of the mysteries that the Square Kilometre Array (SKA) will aim to solve.
Covering an area equivalent to around one million square metres, or one square kilometre, SKA will comprise of hundreds of thousands of radio antennas in the Karoo desert, South Africa and the Murchison region, Western Australia.
The multi-billion dollar array will be 10 times more sensitive and significantly faster at surveying galaxies than any current radio telescope.
The massive flow of data from the telescope will be processed by supercomputing facilities that have one trillion times the computing power of those that landed men on the Moon.
Phase 1 of SKA’s construction will commence in 2018. The construction will be a collaboration of 500 engineers from 20 different countries around the world.
When operational in 2024, the SKA will generate data rates in excess of the entire world’s internet traffic.
ICRAR used an international consortium of astronomers to conduct a survey with the Janksy-VLA telescope, employing AWS to process the data, and they are now trying to determine how the services will work with a larger system.
“Things are changing quickly – if you do something today, it might be different next week.”
Quinn says cloud systems assist international collaboration by providing all researchers with access to the same data and software. They’re also cost-effective, offering on-demand computing resources where researchers pay for what they use.
A mammoth telescope comprising millions of antennas across Western Australia and Africa, the Square Kilometre Array (SKA) will help astronomers tackle some of the big unanswered questions of the universe.
Vast quantities of data from the telescope, due for completion in 2024, will necessitate heavy-duty computing infrastructure. The output from the Australian part alone, located at the Murchison Radio-astronomy Observatory 800 km north of Perth, will exceed a day’s Australian Internet traffic in less than 20 minutes.
The ‘brain’, called the Science Data Processor (SDP), will manage the capture of raw data at the Pawsey Supercomputing Centre in Perth and the processing and archiving of this data into a form that astronomers around the world can access.
“The main goal of the SDP is to bridge the gap between the telescope and the science,” says Vinsen’s colleague, ICRAR engineer Associate Professor Chen Wu.
ICRAR is part of the international collaboration designing the SDP – itself a multifaceted collection of hardware and software. Split into 10 work packages, the huge project is managed by 21 partners in 18 time zones with a total budget of $48.3 million. ICRAR is leading the Data Layer Work Package that will develop systems to manage the flow and storage of the telescope data.
Industrial joint-funders, such as IBM, Cisco and NVIDIA, have been involved since the project’s conception. Commercialisation of SKA technology is expected to flow naturally from the arrangement.
“There will be a significant return to industry, come what may,” says ICRAR director Professor Peter Quinn.
The SDP project, and Data Layer in particular, involves collaboration with a Chinese collective of universities, research institutes and a company. Two such partners are Tsinghua University in Beijing, who are working on data storage, and Inspur in Guangzhou, a contractor for Tianhe-2, the world’s most powerful supercomputer.
Collaboration on the SDP is part of wider investment by China in the SKA and radio astronomy in general. In a separate project, nestled in a natural bowl of limestone in the Guizhou Province in southern China, the largest single-dish telescope in the world is under construction. The Five hundred-metre Aperture Spherical Telescope (FAST) is due for completion in 2016.
The FAST design was an SKA candidate that missed out, but still promises to be a powerful telescope. ICRAR is working with the Chinese institutions involved in FAST to learn from their experiences.
“We are particularly interested in working with the Chinese on FAST because of its enormous scientific potential, but also as a precursor to the SKA technology,” says Quinn.