>>>One important difference with client/server and non-client/server multi-user applications is that client/server applications cut-down on network traffic...as an example, if you performed a query on a remote table with a non-client/server multiuser app, the entire table and index file would have to pass through the network to the 'client' machine where the query would then be executed....in a true client/server application, the request would be sent to the server app, the server app would process the request and send only the results to the client...so if you were searching a 10Mb file to find 5K of data you would cut-down on network traffic dramatically....
>>
>>Joseph,
>>
>>This does depend though... If your queries routinely return large amounts of data then the benefit of client server diminishes. The same is true is you have lots of bandwidth kicking around.
>>
>>Then again these are all "ifs". I remember a quote from Jean Cretien "if my grandmother had wheels, then she'd be a bus".
>
>Well, how much bandwidth can you have out of one machine? If all your files are on one machine, and 10 users are performing queries, the bottleneck is going to be the NIC...right? And client/server applications will be sending less data through the bottleneck, and will, therefore, be faster.
>
>Joe

I have to agree with Joe on this one. Even a large dataset is only a SUBSET of the databases, so there's still less network traffic.

But to restate what I said yesterday, I love to keep copies of as much 'static' data locally so that way server processing AND network trafficking is is reduced.

"If I had wings, I'd still be obnoxious"