The SDP is the Science Data Processor The role of the SDP is to take the enormous
data rates coming in from the Central Signal Processor and to convert that visibility data, which
does not look like an image of the sky, into maps of the sky at radio waves The role of SDP within SKA can be compared to the brains of the observatory SDP interfaces with the control system of the telescope and basically takes all of the data and produces the data products that the scientists will use to do their science with Because of the sheer scale of the data involved it represents a very unique challenge The SKA is entering a completely different way of doing science, especially in radio astronomy The reason for that is because it’s producing
so much data in the meantime that we can’t keep the original data anymore so we have to reduce it before we send it
out to the scientists What’s groundbreaking about the SDP is the
sheer scale and complexity of the processing that it needs to perform High performance computing was not something
astronomy had been using up until now really so it’s a completely new paradigm for us at least But even if you go to the high performance
computing community they would say to you well that’s really, really hard SDP is pushing the fringes of what’s possible
at the moment in terms of movement of data within compute
systems and processing capabilities The architecture has to be able to evolve
over the 50-year lifetime of the project so we have to design a system that’s highly
scalable and highly extensible so we can add on modules and change the size
of it as time goes by SDP is really the perfect marriage of science
and engineering You need the scientists’ expertise as to
what needs to be accomplished and as to how the final products should look
and what they want But you also need a heavy engineering component
to realise that through software and hardware The SDP consortium has tapped on resources
from all over the world and this has seen the participation of most
of the countries involved in the SKA We’ve had to bring in all sorts of expertise,
so obviously radio astronomy domain expertise but also the expertise from compute data centres, specialists in high performance computing We’ve also had a lot of industry engagement to help us make better decisions in terms
of technology selection and confirming our design methodology The team’s ability to clearly express to us the very detailed technical aspects of their architecture they know this inside and out, it’s clear
they worked together really well and that was very impressive I’d just like to thank everybody for the immense effort they put in to getting us to this stage and for the hard work over the number of years There’s still a lot to do but I’m very
grateful that people have come together and produced this excellent outcome It’s been a great pleasure to work with
all of them We rely on their help to keep building for
real this incredible machine.

Designing the SKA – the Science Data Processor
Tagged on:                                         

One thought on “Designing the SKA – the Science Data Processor

  • May 18, 2019 at 11:20 am

    Vai ser top de linha e esse computador, o bagúi é Loko mermão vai desimbassa a treta do espaço vô dá uma amanjada no parangolé kom firmesa mermo tá ligado, esse papo ke desinbolo é mermo pros brauka fraga.


Leave a Reply

Your email address will not be published. Required fields are marked *