[alger]

In cluster computing, a bunch of similar (or identical) computers are hooked up locally (in the same physical location, directly connected with very high speed connections) to operate as a single computer. The computers that make up the cluster cannot be operated independently as separate computers. A cluster, as far as any software or other computer is concerned, looks like essentially one big computer.

In grid computing, the computers do not have to be in the same physical location and can be operated independently. As far as other computers are concerned each computer on the grid is a distinct computer. Computers on a network have a program on them that allows unused resources (usually processing time and memory) to be used by another computer on the network. The speed of the connections between the computers on the grid are relatively slow (Ethernet speeds) compared to the speed of connections inside each computer, so processing tasks are broken up into independent chunks and sent out to different computers on the grid. When a computer is done with a chunk, it sends the results back to the server.

Roughly on a grid, a server log in to a bunch of computers (the grid), send them data and a program to run, and runs the program on those computers, which sends the data back to the server when its done.

In sum, a cluster is one large computer made up of small, similar computers, just as R.A.I.D. is one large hard disk made up of small hard disks. Whereas a grid is a bunch of computers that make their unused resources available to select computers (often a single server) over a network.

[beag]

The biggest difference is the scale. Cluster computing is a small scale, a bunch of computers at your work. Grid computing is on a big scale, for example computers of users from many countries (e.g. SETI@home). See here: http://www.maxi-pedia.com/Grid+computing+distributed+computing