The Grid: The Internet Obsolete?

[Doubleshadow]

I would love to go digging into the various references of practical applications of this computing system; especially the materials chemistry and molecular biology research, but for now I'm just passing along the article for discussion with people more knowledge than myself.

Is the Grid really a system that will replace the Internet? When would this happen for the average person and would it really be so reliable that alternative methods of storing information would be phased out? Is there a communications infrastructure in place that would support this system.

The Grid

I don't see information being stored solely on-line for a while, and I don't think there will be enough of an investment in updating current hardware for this to be practical for household users for a while, but I'm sure the physical chemists I know would love to get their hands on this.

[ClosetOtaku]

I'm not holding my breath at the moment. Two years ago, I was asked to participate as a speaker at a gathering of molecular biologists, physicists, IT engineers, and industry reps who were part of a regional university research consortium. Their focus: Grid Computing. What I gathered from two days of presentations was (a) Grid Computing relies on a federated architecture; (b) a federated architecture means everyone shares; (c) sharing was what everybody wanted, but everybody was afraid to do it, and (d) to overcome this fear, somebody had to change -- and that somebody meant "somebody else".

Some early uses of the grid show promise, but in a federated architecture, you're robbing Peter to pay Paul -- that is, you are using processing power that somebody else isn't using at the moment. If this thing catches on, everyone will want to use it -- and you won't have spare capacity to make it work. (It works well now because there IS so much spare capacity.)

I remember when people believed 4GLs were going to make all other languages obsolete; when Java was going to kill Windows; when thin clients were going to eliminate the PC. They had their moment of Hype, then faded into relative obscurity -- useful, but not killer. So, I think, of Grid Computing. Federated architectures are unsound, security-wise, and given the nature of the Internet nowadays it is unlikely the Grid will be the playground for anything but a select group of trusted research facilities.

But, of course, I could be wrong. "Federated architecture" has become a buzzword among IT sales and marketing folk, most of whom have no idea what they are talking about. "Grid computing" is the next buzzworthy phrase, and we have already seen limited applications using this on the Internet itself (e.g. the SETI project using idle PCs to review collected stellar signal data). Maybe there's a critical mass here that can make it happen. Maybe.