SoHo Science
In the previous post I talked about the long tail of science as independent scientists (or small groups) working outside the mainstream research organisations. Of course it could be an individual within a large organisation who works in a very specialised domain, with few if any peers. But usually I think about the SoHo (small office — home office) individual, as the “extreme” example. There are lots of good scientists already working this way, perhaps making a living as consultants and advisers, but “cloud computing”, and science in the cloud creates an opportunity for this to expand, perhaps to the point where the majority of scientific “product” comes from this long tail.
When talking about cloud computing, I am simply defining it as a stack of resources, available on demand. It could be free, or it could be paid for, but payment is for what is used, and only when it is used. The lower levels of the stack, the storage and processing are already in the cloud (as with Amazon services) and is the first step in lowering the costs of doing science that is fundamental to enabling the long tail. For example, in my own field of drug discovery, particularly computer-aided molecular design, we use software that requires extensive storage and processing. A couple of years ago I bought a new database server and a rack to put it in. I then added five more servers bought from eBay and a tape back-up unit. Since the rack is in entrance hall of my house, I needed a very smart furniture style one, with venting and sound suppression and ended up spending around £10,000 ($20,000) in total. It would have been a lot more if I’d bought the eBay servers new of course. Then there is all the time I spent as an amateur Linux sysadmin setting it all up and keeping it going. It is not enough resource for what we really want to do, but is often idle because we are not doing anything with it. As far as I can see, if I were to do this in the cloud, with a storage and processing provider I would be spending less than a few hundred (£’s) a year, mainly because I only pay when I use it. What’s more I don’t have anywhere near the same time requirements for maintenance and (although I do get a perverse, masochistic pleasure from it) I wouldn’t need to spend anywhere near as much time on sysadmin, and could do more science. But disk space and processing is only step one in putting science in the cloud. It was applications that drove the growth of the PC and it will be applications that drives the growth of science in the cloud. Putting scientific software in the cloud is what will really enable the growth of the long tail. When we do that, we will also get the benefits from the collaboration opportunities that working on-line offers.
Of course, this doesn’t necessarily apply to the lab work, and is not generally applicable to all areas of science, but the long tail applies to niches, and there are plenty of these for which this could work.
Write a Comment