Understanding Spanning Tree

Terry Slattery
Principal Architect

I’m studying for my CCIE recertification and was reading up on Rapid Spanning Tree (RSTP) operation. Cisco has a really nice white paper on its operation, with a lot of good pictures and excellent description of how it works and how it interoperates with the older 802.1d STP protocol: Understanding Rapid Spanning Tree Protocol (802.1w)

RSTP automatically falls back to the 802.1d operating mode when it encounters the older switches. This implies that important VLANs that need fast convergence should be built using newer switches that support RSTP. Move the older switches that you still want to use to less critical parts of your network.

There’s also a nice white paper on 802.1s, Multiple Instance Spanning Tree (MIST): Understanding Multiple Spanning Tree Protocol (802.1s)

MIST relies on RSTP internally, so it is useful to also be aware of how RSTP works. The MIST document doesn’t go into the level of detail that the RSTP document does in describing the transactions between switches running the protocol.
There are a number of potential configuration problems when MIST is used with other switches that run a non-MIST protocol (PVST, RSTP). Because of these potential problems, if you’re planning a switch to MIST, make sure all switches in the spanning tree run it. The best way to have a stable network is to avoid potential problems.

In thinking about where I might want to use MIST, I thought of the Cisco Solutions Reference Network Design (SRND) guides. The current Cisco thinking (I’m not up to speed on other vendor recommendations – leave a comment if you know what they are) is to do routing at the access layer, or certainly at the distribution layer. The result is that VLANs would be constrained to a wiring closet (think subnet per closet). This should limit the number of VLANs handled by a given topology and the limited topology would create more stable VLANs (fewer places for trouble to occur and smaller spanning tree domains make them easier to troubleshoot when they do occur). So, why would you implement MIST? Please leave a comment if you’ve implemented it and are willing to describe why.



Re-posted with Permission 

NetCraftsmen would like to acknowledge Infoblox for their permission to re-post this article which originally appeared in the Applied Infrastructure blog under http://www.infoblox.com/en/communities/blogs.html


Leave a Reply