An artist's impression of the Square Kilometre Array's low frequency aperture arrays

by Economics Correspondent Seán Whelan @seanwhelanRTE

No doubt we could all think of better things to do with €3.1 billion than paying off the promissory notes used to bail out Anglo Irish Bank and Irish Nationwide Building Society.

My suggestion is we pay for the Square Kilometre Array.

The SKA, as its known, is a project to build the world’s biggest radio telescope, with a receiving area of – yes, you guessed – one square kilometre.  Not that it’s going to be one humongous version of a sky dish.  The idea is to scatter several thousand antennae over a very large area.  By spreading your one square kilometre of receivers over a radius of 3,000 kilometres, you get the ability to collect vast quantities of very accurate data.

Among other things, you could produce the first 3-D map of the observable universe, taking in an estimated one billion galaxies.  And apparently you could get all that for just one year’s instalment payment of the promissory note.

The SKA consortium (a ten nation group headquartered at Jodrell Bank, part of the University of Manchester) presented at the ESOF science convention in Dublin last summer.  They said the initial build cost estimate for the array was €2.5 billion.  Of course projects like this always overrun, so lets bump up the figure to …oh, why not  €3.1 billion, for the sake of argument.

Last spring the Consortium decided where it is going to build the array.  It will be split between Southern Africa and Australia (giving the 3,000 km spread).  Most of the antennae will be in Southern Africa, centred on the Karoo desert area of the northern cape in South Africa, but spreading out as far as Ghana and Kenya.

The Africans are very keen on the project, not just as a big project in its own right, providing employment and career opportunities, but for its value in inspiring African youth to follow science and science related education.  Few things in science capture the imagination like astronomy, and the South African government in particular is placing a lot of emphasis on raising the horizons of of African youth and inspiring them to imagine what is possible in their own continent and in their own lives.

The SKA is already rolling out as a teaching aid in some African schools.  It could be a step change in thinking about development for the region.

But it’s not just African schoolkids that are inspired by the project.  Big business, in the form of IBM, Microsoft and Siemens Nokia – to name but a few – are very interested in the project because of what it means for big data.

According to the promoters of the SKA, when the full array is switched on sometime in 2024, it will capture as much data in its first two hours as currently exists on the internet.

Yes, go back and read that line again.  It is a truly vast amount of data.  Which brings with it all sorts of practical issues such as how and where do you store it all?  More pertinently, how do you process it?  And particularly for fans of the broadband/fraudband debate, what kind of network capabilities do you need to move such vast amounts of data from remote collecting sites (and the more remote the better, for data collecting purposes), to central processing centres?

I listened to a pointy head from Nokia Siemens explain, in terms I couldn’t possibly understand, the engineering challenge involved in shifting data around the SKA project, and how meeting that challenge could help his company bring products to market faster that should make things better for the rest of us who use (and pay for) telecoms networks.

The key to it all is exaflop computing – the current holy grail of computer technology.  This is a machine that can carry out one quintillion flops – that is one billion billion floating point operations per second  (or 10 to the power of Eighteen operations per second).  Current estimates say the earliest we will have such computing power will be 2018 – assuming some pretty daunting challenges around cooling, energy consumption and the physics of microprocessors are cracked.

Computers of this power will have huge implications for the way massive quantities of data is processed, from global weather forecasting to economic modelling, not to mention sound and video files uploaded to the net.  Bank of England guru Andrew Haldane has spoken of his dream of a kind of star trek room, where all the financial transactions in the world could be viewed by regulators on a real time basis, so they could see where money is flowing.  You probably need some exaflop scale computers for that sort of thing.  The SKA buildout – between 2016 and 2024 – is the kind of deadline the pointy heads like to work towards to try and get these problems solved.

India is set to become the eleventh member of the SKA consortium.  New Zealand is the smallest country involved, Sweden the next.  Membership costs €1m for the first four years, before contributions for the buildout phase are sought.  Right now it seems Ireland cannot afford the €1m it costs to buy into the project, which promises to throw off a lot of valuable intellectual property, directly and indirectly – as well as the workshare arrangement common to these kind of projects, which keeps a lot of money in consortium member states.  Yet for the price of one promissory note instalment, we could build the SKA single handed.

So to sum up, you could spend €3.1 billion building the biggest scientific apparatus on the planet, stress test Einstein’s theory of general relativity,  kick starting a new kind of development in Africa, boosting and focussing research and development of the kind of computing power and network speeds that the emerging digital economy demands, establishing a new economic driver and creating thousands of high end downstream jobs, and acquiring vast quantities of knowledge on behalf of all human kind.

Or you could, as usual, just e-mail it to Professor Honohan to delete from the ELA account on behalf of Super Mario and Co.

Makes you think.  Makes you sick.