Slashdot Mirror
Visions Of The Future Of Grid Computing
CaptianGrid writes "Computing grids, or software engines that pool together and manage resources from isolated systems to form a new type of low-cost supercomputer, have finally come of age. BetaNews sat down with some of the world's leading grid gurus to discuss the significance of such distributed technologies and separate grid hype from grid reality."
Then more recently we have seen Univa being created, which I am involved as founder and advisor
Univa
Univac - a successor of Multivac, the largest computer in Asimovs world.
Nerds - they get everywhere.
Check out Apple's X-Grid technology!
It runs on any OSX system, 10.2.8 and up. Put your spare cycles to work.
Xgrid: High Performance Computing for the Rest of Us
Never ask for directions from a two-headed tourist! -Big Bird
The article mentions the commoditization of grid computing by adhering to a set of standards, but past a certain point, it makes little sense for IBM or Sun to make their tools interoperable... that makes their consulting value-add on top of grid resources they offer diminish.
I think that for full standards compliance, you'll need to look to companies which don't offer their own computing resources -- platform-agnostic companies. But then who do you buy the compute resources from? Unless you're buying your own systems for use (which makes "utility computing" less viable), it's a bit of a catch-22.
500GB of disk, 5TB of transfer, $5.95/mo
If you want examples of operating systems that help with gridding, check out Plan 9 from Bell Labs and it's sister project Inferno. Nice thing about Inferno is that it runs on Linux, Windows, Mac OS X, Plan 9, and on native hardware.
Free of Flash! Free of Flash!
There will never be a substitute for a single box with a lot of CPUs on it. For tightly coupled dataset the latency of a grid will be a limitation.
Transcend Humanity. Please.
What we provide is primarily an implementation of Web services standards to allow people to build services, and the primary goal is also for us to provide a set of pre-defined services that allow you to use Web services protocols to interact to request the allocation of compute resources, the creation of computational services and moving the data from one place to another and so forth.
Does this sound like Carly Fiorina attempting to explain HP's strategy to anyone else?
If you've got a problem that's trivially parallelizable, then sure grid computing is great! RC5, seti@home, and similar projects can benefit from grid computing (really, that's what grid computing is -- someone else's code able to run on your machine when it's idle and do work).
However, don't even begin to think you'll be solving anything that requires any sort of processor to processor communication. Rocket simulation (our local favorite example here at UIUC) for instance is heavily communication based.
The linpack benchmark that top500 uses also needs a low-latency interconnect to perform really well, so don't expect to see "the grid" sitting up at the #1 supercomputer slot on top500.org anytime soon (or really, ever, unless someone develops FTL networking). Latency on the internet in general (and specifically around the world thru all those switches and latest_slashdot_hot_chick_movie.torrent packets) is nothing near what a supercomputer needs.
Now, there are research groups looking at ways of making communiation delays less of a problem, including the one I was in while I was in grad school. There's a number of ways to do it, but none of them I've seen are going to take on worldwide-network-latency and survive with their performance intact.
Even something as "simple" as chess wants to have a fast interconnect - every node that's gotten stranded working on low-priority (bad move) work is a wasted node you may as well not have.
Slashdot Patriotism: We Support our Dupes!
What no one is mentioning is that these big cluster/grids that Sun/IBM are building to later sell over the network are dependant on the ratio between network speed and batch file sizes.
EXAMPLE: IBM is currently offering CPU/Hour service in Houston to oil and gas companies. Sounds great till you realize the multi-terabyte files that consume such a massive compute service are too big to be readily sent over the network. Instead they use vans to haul tape and disk over to IBM and then run the process on it.
What is the bandwith of a station wagon? Right now its faster than the internet on a 20 mile drive across Houston.
But even take it a step further and the ratio remains. What if I wanted to pay Sun/hr for CPUs while I worked on a big Maya render of 200 gigs. By the time I've sent that over cable modem have I gained a ton in performance time?
The problem I see is that we are making CPU massively parrallel but not networks. So will it EVER make sense to send a massive file to a commercial grid over a singular network connection.
Somone should do the math.