Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. If you'd prefer to explore, try our test area to get started. And see here for current known issues.

New Member

Why is this wrong?

Hi everyone, below on the left is a quick diagram of a section of our network. Our core switches are also routers and use HSRP. The core and distribution switches have about 150 VLAN's on them, even VLANs on the B switches and odd VLANs on the A switches. But only one VLAN on the access switches, and there is no trunk between the distribution and access switches.

The boxes behind the switches represent seperate buildings. There are only 2 usable pairs of fibre between the distribution building and access building.

Our current setup has redundant links from access switch A to the two distribution switches. If the access A switch fails, the B and C switch aren't accesible.

Now my proposed solution.....on the right. I've mentioned it at work and I get shot down. Appearently we had a CCIE at work once and he said it was a bad idea creating a box like that.

The way I see it STP will just block one of the two links going out to the access switches. If access switch A fails then the other two still stay up, as opposed to the current setup.

So what is wrong with my proposed solution?

Basically my boss wants solid proof why this isn't bad, but I can't seem to find any Cisco recommended material stating that it's good or bad.

  • Other Network Infrastructure Subjects
5 REPLIES
Purple

Re: Why is this wrong?

Hi,

I don't see any issues with your design. The purists will argue that the correct way to design such networks is to dual-home your access-layer switches to the distribution-layer switches and the distribution-layer switches to your core switches. However, we cannot always have that luxury. In terms of whether it will work or not, you are correct that STP is not going to be an issue.

I don't believe you will find any documentation explicitly covering this topology (since it is not completely in line with the hierarchical design model). However, that does not mean that this model will not work. And the alternative (your present setup) is far worse..

Hope that helps - pls rate the post if it does.

Paresh

New Member

Re: Why is this wrong?

Thanks for your reply, although I still am curious why the CCIE would have recommended the current setup versus my proposed setup..... I can't make sense of that.

Purple

Re: Why is this wrong?

I'm curious too... But then, not everything a CCIE says has to be right !

Paresh.

New Member

Re: Why is this wrong?

Ok, just one variance...we have this current setup in a few buildings, and in a couple buildings, as the access switch A, it's a 4000 switch, with 2950's as the B and C switch. Would linking the redundant link from the distribution B to the access C switch still be acceptable?

So if lets say distribution A switch went down, all the traffic on the 4000 A switch would be daisy chained though the B and C switch back to the distribution B switch.

New Member

Re: Why is this wrong?

Nelsonpapel,

I think you may have to figure out what the requirements for the orange building are.

I agree that spanning tree will just work it all out. You might want to look into Rapid Spanning tree for a while to see if that fits your requirements.

I can see why if the orange building isn't that important to your organisation why you wouldn't bother futher complicating the network.

Right now people in the orange building are complaing it's most likely Access A. complicate the nework by bringing C into the picture. well you get the drift I'm sure.

Tony Henry

94
Views
0
Helpful
5
Replies
This widget could not be displayed.