You are here

Blue Collar Computing

Leave it to Ohio to apply the term "Blue Collar" to information technology. The Ohio Supercomputer Center has a program called Blue Collar Computing. Some quotes about it:

The complexity of current HPC [High Performance Computing] hardware and software creates a substantial "barrier to entry" for both scientists and engineers. Without proper HPC tools many of our modern research problems range from extremely difficult to impossible to solve. In short, we are losing opportunities for innovation due to an incomplete national HPC infrastructure.

The most formidable barrier of HPC adoption is the lack of simple and cost-effective tools available for use. Just as the graphical user interface (GUI) made desktop computing accessible, and web browsers made networking popular, the right tools are needed to make HPC widely effective.

That is, supercomputers are too hard to use for most businesses.

OSC lists some good next steps to make Blue Collar Computing happen, like public-private collaborations, training and better tools, but one focus is clearly missing in my view: focus on the total user experience by utilizing user-centered design methods. A focus on reducing complexity, usability, understanding user needs, iterative design, etc. is what has made the GUI and the web what they are today. It is long overdue for supercomputers to catch up.

Writing at Clustermonkey, Stanley Ahalt and Kathryn Kelley (from OSC) sum it up well in their article HPC for the Rest of Us. The biggest barrier is: "Hard to use means hardly used – at least by the broader community".

The other news is that this may be going national. USACM reports that legislation for a similar effort has been introduced. About the bill:

We shall see if this catches on at the federal level. And if "blue collar" joins "user friendly" as another way to say "easy to use".

Comments

It makes sense. Having basic tools to make the job easier would definitely allow for greater use of a technology. I don't think I would have named it like they did but I think that they have the right idea with wanting to create tools to make super computing more accessible. I wonder if the super computer at OSU is being under utilized?

Actually, quite the contrary, systems at the Ohio Computer Center are currently at peak levels (>90%) and there is a long queue line to run on them. Scientists, researchers, and large corporations such as proctor and gamble are the typical users of Supercomputing technologies. This is due to the fact of the cost of expertise to keep them running and the knowledge to run experiments and software on them.

The point is to make CEOs and politicians to invest more money into R&D and computing as there is a large return on investment. In turn it helps out the little guy.