I am trying to understand the pros and cons of the two differing designs for an access layer with dual L3 switches acting as collapsed core\distribution.
If we have two layer 3 switches DIST1 and DIST2. Both running HSRP with DIST1 the active for all vlans and also configured as the root bridge. DIST2 is the standby also is also the secondary root bridge.
At the access layer in one of the remote data clsoets we have 3 Cisco 2960's (not my choice but this what I have). There are 2 uplinks available from the remote closet to DIST1\2.
My question is is it better to have one of the 2960's act as the uplink to DIST1 and DIST2 and chain the 2nd and 3rd 2960's from the first OR have the 1st 2960 uplink to DIST1, the 2nd 2960 connects to the 1st and 3rd, and the 3rd connects to the 2nd and uplinks to DIST 2.
So Option 1 looks like this
2nd 2960 3rd 2960
Option 2 is like this
1st 2960---2nd 2960---3rd 2960
Option 2 provides the benefit that if any of the access layer switches fails then the others could still connect to the core.However, it also means that the 3rd 2960 will have use the path via DIST2 as the STP path cost is less than going via 2nd, 1st then DIST1. I suppose we could modify the path cost on the uplink from 3rd to DIST2 but I am not sure if this a good idea? Or what other implications I would need to be aware of?
Option 1 is simplier and we dont have the STP path issues, but it also means that 1st 2960 is a single point of failure.
I would be interested to know what others think and how you might have addressed this in your own networks?