Alright, okay, and let's see if this goes forwards.

So again, I'm Juan de Santander Verna, I'm currently the, what we call the lead product

manager for SK software, so I'm responsible for essentially creating the roadmap for the

software that we were doing. And this is the outline that I'm following for this talk.

First, a bit of an introduction to the Square Kilometre Array, then software, which is at

the core of the SK Observatory, why we needed to scale our software development, why we

chose SAFE for that, and some future work that we still need to do. So why do we need

one square kilometre of collecting area, and can we find a radio-quiet place for it? This

is a question that puts the SK, the SK, we built in it, in this realm of 21st century

observatories. I've tried to align them with the electromagnetic spectrum, but of course,

there are some of these observatories, even not observing the electromagnetic spectrum.

They might be using the electromagnetic spectrum to derive some information, but the Square

Kilometre Array, when it's fully constructed in 2027, what we call the phase one, will

be one of the largest, large-scale infrastructures in the world. And the idea is that we're going

to be studying the H1 to enable the SK science from testing general relativity through pulsar

time ensembles, the first stars and galaxies, cosmic magnetism, and essentially brings what

we believe is one of the broadest ranges of science from the facility worldwide. And for

doing that, we have two technologies of telescopes. The first one, you saw more traditional dishes,

and the other ones are more type of kinds of antennas. They will work as two world-class

telescopes, but working as a world-leading observatory, you will be able to submit proposals

to both of those telescopes at the same time. And they are based on two of the most radio

quiet zones in the world. Western Australia, we have the Sk1 Low, and in the Karuba Desert,

in South Africa, we have the Sk1 Mid. So these have very stringent radio-quiet zone

regulations. We will not be able to do the Sk1 if we're not for that. And we have to

be operational for 50 years, which is quite a stretch. These are the current make-up of

the SK organization, with 14 countries included. And you can see that we are still in discussions

with several members. Switzerland is very close. And we are in the process of becoming

an intergovernmental organization. And for that, we already have the treaty signed in

Rome on the 4th of March, 2019. And those three countries with the check marks, Italy,

South Africa, and the Netherlands, have already gone through all the ratification for it.

So we would go. We still need Australia and the United Kingdom. And then we will have

the intergovernmental organization. And we passed this TBR in December 2019. So all of

the previous phases of design and architectural analysis and that, they are behind us. But

what we did is to start working on this software, which is at the core of the SK observatory.

So, if you've seen this picture before, the funny thing is that nowadays everything is

digital. And it's funny because it's true. So for the SK-1 observatory, we have a lot,

we have already 100 FTEs per year during the bridging phase. So this is not even construction.

We will have just that number of FTEs for the data processing part of the telescope.

And the 50% of the total manpower, sorry, total effort is for software. So this is a

remarkable amount of effort that we are putting into the software. If we look at the SK as

a combined system, SK-1 low on top, SK-1 mid below, we have two main pieces of software,

which are what we call the science data processors. These will be computers of around 130 petaflops.

And they will have to do all of the processing, all of the data ingest, all of the integration

with the relevant metadata, but also with a correlator, we'll have to keep a real time

calibration flow that goes on with it. So you also need to be computing those real time

calibrations. And we will be generating in the other 300 petabytes per year per telescope.

That will go to what we call the SK regional centers. There will be a multiplicity of them.

And each one is a supercomputer on its own. So we also need to write the software for

these machines, which will be largely studied from the science data processor software.

And we need all of the control software that goes with it. So there is a lot of software