For the project I am on, we need to put a very fault tolerant network in place with very high available throughput. There won't be very much traffic on the network, but one of the keys to the success of the project is to demonstrate a solution that can scale to very high traffic loads.
My first pass is this:
- Two physically separate server farms with four servers in each.
- Two physically separate workgroups with four workstations in each.
Use two Catalyst 5505 switches as the core with Gigabit ethernet line cards.
Use one Catalyst 3524 switch in each of the four "areas" configured to use Gigabit Ethernet to uplink to both of the 5500s and 100baseT to connect to the servers/workstations.
This should provide a redundant path from each "area" to every other "area". The Spanning Tree should then detect a core switch or link failure and provide a secondary data path.
To make the design even more redundant I could add a second switch in each "area" and put two NICs in each computer and connect them to both switches.
I apologise for using "area", but just couldn't think of a good word to use instead. I think this design works, but would like someone to agree or disagree (or propose a better one).
thanks,
bruce
