Take a network of 2 servers
1 - Windows 2000 - file/print and user management
2 - Windows 2000 SQL Server - database only
and 20 workstations of which 10 run an app that talks to the SQL server.
Network is TCP/IP and is hub-based (rather than co-ax).
With that setup ...
1 - Am I right in thinking that every packet will visit every machine on the network?
If yes, what decides the sequence of the "visits"?
Can the physical location of the servers make a difference to performance? i.e. which socket on the hub the servers are connected to?
Now add in a WAN link - with no filtering.
Would all the packets from both networks now visit every pc on both networks?
If so, what controls the sequence.
Now add filtering so that packets for THIS network do not go across the link to THAT network.
When a packet from THIS network is generated and DOES need to go across to THAT network, when it reaches the other side, will it visit all the pcs?
I've been asked to "speed" up the application. I'm a programmer, not a network specialist. Ideally, if it is just a case of some settings on what gets sent where, then I can probably manage that.
Windows networking with (as yet) an unknown link between 2 networks. No idea on what filtering orrouting is installed.