Mike W., I appreciate your comments. They did indeed have spanning tree problem in the beginning. An older switch won the root bridge election because of its lowest MAC address and was having an issue with its uplink. Once they called me into the problem, we moved Root Bridge to the core of the network on a more robust switch. We used a bridge priority of 1, 2 & 3 for the switches closest to the core.
I can appreciate your recommendations on bandwidth vs. switch hops, that was my belief, but I can't figure out why they have so many problems. Because the student's traffic goes about 99% to the Internet, their bandwidth requirements don't amount to much. The Gig links in the network are less than 10% utilized (most around 1%) and the 100M links are about the same. They don't want to pay for a full network audit, but they always lay the blame on the network hardware, even through there's no indication the equipment is mal-functioning. That's why I am going down the garden path on the design of the network as being too deep in switch hops. Again, I appreciate your opinion, it makes sense! Jeffrey Reed Classic Networking, Inc. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Michael Williams Sent: Sunday, March 17, 2002 1:09 PM To: [EMAIL PROTECTED] Subject: RE: Design Best Practices [7:38453] Although I agree that three "Layer 2 hops" is best, as Cisco recommends (core, dist, access), I also can't understand why the scenario you describe above is causing a problem (assuming the spantree roots are set properly, etc. be careful that one of the 1900s isn't the root bridge for a VLAN =) I don't really see the number of Layer 2 hops as being a problem. About the only problem I could see is if one of the switches along the line is getting overloaded with traffic. Given the information you provided, there could be a couple of "weak links". If one assumes the max throughput of a hub is around 4Mbps, then each 1900 would have 4-8 * 4Mbps = 16 - 32Mbps on each 1900. Each 3500 has 8-12 1900s (let's say 10), so each 3500 has 160-320Mbps, no problem. But then Each 5000 has 10-12 3500s (again, let's say 10), which is 1.6-3.2Gbps. This could be a problem for the 5000 and it's 1.2Gbps backplane. That's weak link #1. Also, assuming the 5000 isn't being overloaded, even being generous, the 100Mbps links between the 5000s and 5500s are definitely a weak link. At that point you have the equivalent of roughly 600 (10 3500s * 10 1900s * 6 hubs = 600) 10Mbps hubs trying to jam through a single 100Mbps link. Even if each hub os only using .3Mbps (less than DSL/Cable modem speeds), that's still almost 200Mbps trying to jam through 100Mbps link from the 5000 to the 5500 (assuming most of the traffic is destined for the internet, or a destination being routed by the 6500(r), etc) IMHO, those two "weakness" are 100 times more likely to be causing problems than the number of Layer 2 hops involved, again assuming STP is configured correctly. Once the 5500s uplink into the 6500s, it's virtually impossible for those 5500s to overload the 6500 (and it's default 32Gbps backplane), especially since (according to our Cisco rep) the Gig ports on the 5500 deliver no where near actual 1Gbps performance. So, in a nutshell, IMHO, bandwidth, not the # of layer 2 hops, is more likely causing your problems. Just my 2 cents. =) Mike W. Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=38610&t=38453 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
