Well, basically distributed computing is dividing large processess to multiple computers. Unlike clusters, DCs make use of un-used idel computing powers available from around the whole people like you and me.
A distributed server will assign work units to clients who request it. The client will calculate it at its own pace depending on its free time. Then will return the final result to the server and server take the record and assign another work unit.
For more info, check out here.