Slashdot Mirror
Amazon Betas 'Elastic' Grid Computing Service
RebornData writes "I receieved an e-mail this morning inviting me to participate in a limited beta of Amazon EC2: the Amazon Elastic Compute Cloud. It's a grid computing service that allows you to create and upload your own Linux-based machine images and run them in Amazon's system, starting at $.10 per "instance hour" (each machine instance being equivalent to a 1.7GHz Xeon with 1.75GB of RAM, and 160GB disk). You can use their tools to create and start new instances dynamically to meet whatever your particular capacity needs are at any given moment. Fedora Core 3 and 4 are explicitly supported, but any distro based on the 2.6 kernel should work. The service documentation provides more technical details. Unfortunately, it appears that the beta is limited to existing Amazon S3 users, and is already full."
Compromised Amazon grid being used as a botnet to send spam. Letting people upload their own OS image is really asking for it.
ELOI, ELOI, LAMA SABACHTHANI!?
what? So they are building computers for you and letting you use them for whatever you desire? Hmm. 0.10/hr thats $2.40/day. Thats $876/year. Not a bad deal. Heck its even cheaper than some website services that give you a dedicated server for $200/month.
You can also go fart around with Amazon's Web Services for fun and profit.
They rolled this out a few months back, when I was one of the brave few to sit through the presentation at a programmers conference in Santa Clara (for a free t-shirt and pen.) It was actually amazingly cool and I'm planning something of my own with it. (but I ain't telling you because I wouldn't want anyone tempted to swipe my neat idea, like thinkgeek did to me once already.)
A feeling of having made the same mistake before: Deja Foobar
Perhaps I have the economics wrong, but isn't it more cost effective to build your own cluster out of discarded PCs?
You've got the economics wrong. Building your own cluster out of discarded PCs is not economic. Building your own cluster out of brand new PCs might be.
Still, leasing is attractive for many reasons. Such as predictable costs, complying with yearly budgets, etc... If you build your own cluster, and find it doesn't work as expected, or you didn't really need it, or whatever, you are pretty much fucked. Clusterfucked!
It seems like the only serious problem would be getting the 20GB TIFF (or 8 GBs of WAV files) over to the server instance in the first place.
Having to move all your data over to the server and back adds significant set-up overhead, particularly if you only need the monster for 2 hours at a time. When you need the numbers crunched on demand, you don't want to have to wait 6 hours while the data set squeezes its way to the bay area and back over routers.
I'm sure that there are applications for this, but quick-turnaround stuff is hamstrung by the I/O bottleneck known as the Internet.
Don't blame me, I voted for Baltar.
I wonder if there will eventually be a dependent subculture built around grids. Companies who sell pre-written (and customizable) solutions that require a grid to run. Companies might buy/license them and run them quarterly, to (say) update/analyze their data warehouse. Or if somebody develops a "killer app" for grids, then they might set up their own and rent it as a distributed computing solution. Maybe develop some kind of adapter for Mathematica, let people run their huge econ simulations, then truck their results away to slice & dice.
Just junk food for thought...
You might want to check out Starfish. It's Google's MapReduce implemented in Ruby, kind of. It makes distributed grid computing possible in six lines of code. Unbelievable, but true.
Ian
PS I've personally got nothing to do with Starfish. I read the author's blog--that's it.