I take it you mean Shielded Twisted Pair eg Cat5 etc.
Depends. I use Cat 5e and try to use Cat 6 where possible on Gigabit installations. Also bear in mind that out of the box, Gigabit is "only" about 5x faster than 100Base. To get closer to the real theoretical 1000Base you need to enable Jumbo Frames, which is not supported on all adapters (Intel seem pretty good in this respect). It's a similar concept to tweaking MTU, or, the difference between two trains travelling at 100 mph but one of them with all carriages full while the other is only half-occupied. Same line speed, different throughput.talking of cat5, we just upgraded to a managed gig switch in our server room because a large percentage of our Dell machines are also gigabit ready, a couple of client machines refuse to run at full speed in certain locations [auto negotiating at 100 full duplex] so was wondering, if older cat5 cable was used would it be the cause, or do all cat5 cables allow gigabit speeds ?
most of the wiring was installed behind partition walls when the factory was built 12 years ago, so signal path is :-
Server > Gig Switch > Patch Box[cables run from patch box thru wall to wall sockets] > Wall Socket > Client PC
Oh yeah forgot about that, bloody acronyms, anyway we don't run spanning tree, but one of my objectives for the year is to produce a report on how we could implememt it. At the moment we have about 180 switches in a flat network config. Spanning tree could be a PITA to implement. Usually loops happen at the fringe of the network (in my experience) apart from today [:$]. So have been thinking about implementing spanning tree on the edge switches and then bringing it into the core eventually.