Poster Abstract

P8.1 Slava Kitaeff (The International Centre for Radio Astronomy Research)

Theme: Telescope operations and scheduling: from classical to autonomous

Prototyping Australian SKA Regional Centre

The first phase of the Square Kilometre Array (SKA) Observatory will include two telescopes: SKA1-Low to be built in Australia, and SKA1-Mid in South Africa. By the time the SKA is in steady-state operations in the late 2020s, it will be generating raw data at the rate close to 1 TB/s. After initial processing, the output data products to be disseminated to science teams are estimated to be up to 600 PB per year for the two arrays combined.
The core construction funding for SKA1 does not include the distribution of data to users, nor computational facilities to enable users to undertake further data analysis, both of which are necessary if the SKA is to deliver on its scientific promise. As with several other large-scale scientific instruments, these functions will be provided from regional funding rather than from project funds. The SKAO has introduced a model for SKA Regional Centres (SRCs), which would see SKA members establish and fund centres to provide computer support for the astronomers in their country or region.
Australia has built two SKA pathfinders - ASKAP and MWA. Significant effort and investment have already been made that paves the way to the future Australian SRC, including over 35 PB of data produced by MWA and ASKAP and stored in the Pawsey Supercomputing Centre. Australia has formed a partnership between several research organisations to develop a roadmap towards the SRC in Australia, starting with tools needed for the SKA precursors and aiming build-up towards full SKA operations.
We present a prototype of a system that tests a number of assumptions and requirements to Australian SRC to post-process and analyse multi-TB and multi-PB radio astronomy datasets.