I've just received my shiny, new, brushed steel and polished glass smartphone. I'll admit now: I'm in love with it. Before this day came, I'd talked frequently about how amazing life would be after it arrived and, now that it's here, I spend a lot of my spare time bathing in its visual modern beauty and trying to fully realise what I'm convinced are life-enriching possibilities.
One of the important features of this smartphone is the loudly-touted easy access to a cloud on which data can be stored and complex calculations can be made.
In 1950, Herb Grosch, a Canadian-born astrophysicist who worked on the Manhattan Project, envisioned a future where the whole world would run using a cloud system of computing, operated by individual local terminals but served by only 15 data centres.
Today, there are about 50,000 data centres just in Australia – certainly more than 15 worldwide – but they are getting larger in size and smaller in number.
Arguably the largest data centre in the world, the Lakeside Technology Center, is 10 hectares (24 acres) in size. It's the nerve centre for Chicago's commodity markets and requires about 100MW of power to operate.
The architecture of computers means they can't presently solve certain kinds of equations, but what they can do is break a model down into millions of tiny parts and approximate the solution to governing algorithms by iterative methods. The smaller the parts the problem is broken into, the greater the precision and accuracy of the solution, but the more calculations required.
My final year project for my Engineering degree, an aeon (or ten years) ago, was to design an efficient shape for a solar-powered car, hypothetically to be built and raced in the World Solar Challenge. It involved modelling the movement of air across the surface of the car, a problem governed by partial differential equations, unsolvable directly but susceptible to a good approximation. My team would go into the computer rooms on campus, enter the surface geometry of a car we thought would slip through the air cleanly, and then leave the post-processor to think about the problem, which would take about a day. The graphic visualisation part of the problem would also take hours. We would return, review the results, think about how shape could be improved and do it again.
How much quicker would the process have been an aeon later? With the accessibility of clouds now, computational fluid dynamics packages and so many other data analysis packages for professionals from structure designers and advertisers to baseball scouts can be operated remotely, by our smartphones or tablets, which need only have the power to display the interface between the calculator and the user: a "dumb" terminal.
Computing is rapidly becoming a service business. Want to store your precious data? Don't keep it where moth and rust destroy. Leave it all with us for a monthly fee, or for free if you promise to notice our constant but subtly-placed advertising banners.
Need a complex problem solved? It was not unusual, until recently, for a seat with a data consultant to be up to tens of thousands of dollars. The barriers to entry are now lower for modelling and data manipulation consultants, such that all it takes is a short lease contract for software and the computer power on which to run it (and soon, your sexy smartphone).
The challenge for data centres is business continuity delivered in an efficient way. The global ICT industry was estimated in 2007 to be producing 2% of the world's carbon emissions and data centres 14% of that, the latter of which appears to be growing. Google keeps the server hallways in its centre at 27 degrees celsius to reduce airconditioning loads. Other centres are being built near proposed tidal power generation sites, such as one near the Pentland Firth in Scotland.
The technology can be used across the entire spectrum from trivial to world-changing problems. There are, for example, teams of people involved in the search for intelligent life beyond Earth. The SETI@home project used hundreds of thousands of idle home computers to review reams of data from a radio telescope array for faraway signals that couldn't be dismissed as noise. Last year, Amazon donated a part of their cloud so that SETI could continue their efforts with even greater power for the next six years.
It's frustrating but also amazing that the problems we want to solve seem to become more complex the more we learn. The cloud will no doubt become the way that we will relate to and get closer to solutions to the most tricky and long-standing unknowns.
 |
No comments:
Post a Comment