Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

A. Is using one network switch is faster vs 3-4 network switches connected together in same room?

B. Is there difference in speed at all between cheap switch and more expensive model?

P.S. network with 30 computers and 30 ip-phones.

share|improve this question

A. It's not going to make anything any faster to separate workstations into multiple switches mainly because when you do that, you will be forcing interswitch traffic through an uplink between switches which could turn out to be a bottleneck. Usually this would be a single gig uplink but even if you bundle a few ports together in a channel it will still be shared amongst everyone connected to that switch.

B. You can't just go by interface speed. They will all say 10/100/1000Mbps. Different switches have different backplane switching fabrics. Cisco switches usually (but not always) have a fabric capable of supporting a near full switch load. See here for specs on many Cisco switches. You can google similar specs of other switches for comparison but I would bet the cheaper you go, the worse it gets.

Specs on a couple HP switches see "performance" on the third page here and here

share|improve this answer
That's a very good point for B. I've seen 48-port gigabit switches that can only run at about 12-ports full duplex before their backplane is fully saturated. – Mark Henderson Dec 6 '11 at 1:31
Thanks, I mentioned it because when running a phone and PC off a single port, the utilization per port tends to be a bit higher than data only especially in a call-center environment. – Paul Ackerman Dec 6 '11 at 1:33
I have separate ports though for phone and pc – barcuda555 Dec 6 '11 at 2:05
If you go the multiple-switch route, it's probably better to arrange them in a star pattern rather than daisy-chain. Then you have a maximum of four hops between endpoints (device-switch-switch-switch-device) and you decrease the chances of saturating an uplink between switches. If you had four switches daisy-chained, you would have 5 hops. – Jonathan J Dec 6 '11 at 4:36

My answers

A. Yes, although it likely doesn't matter with only 60 devices. One good switch will be better than 3-4 cheap switches. 3-4 small good switches will out-perform 1 cheap bigger switch if the network is properly arranged.

B. Heavens yes! The backplane speed (aka "fabric" speed) and throughput are dramatically different among classes of switches.

My advice

Get good switching, and setup separate VLans for voice and data. Then setup the phone system and DHCP to put the phones onto the phone VLans.

Cisco is the gold standard, although it is expensive to buy and more expensive for maintenance. I like HP ProCurve. Juniper is also good, although the configuration can be difficult.

share|improve this answer

A. It can be faster if you have a lot of traffic between the workstations that are each connected to the different switches. You're limited by the link between the switches.

B. This is true, but also only if the ports are heavily used.

One thing I feel it's important to mention is that if you get more than one switch and one dies, you haven't lost your entire network, only the devices on that one switch.

Lastly, more expensive switches have more options/functions -- you can segment traffic into different virtual networks. This is especially important with VOIP traffic when you need to have a certain level of quality. You can mark VOIP traffic as more important, so your phone users are less likely to have call problems in case the computers use too much bandwidth.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.