Clarification on MST + CIST theory, need help

Mrock4

As a reference, I'm getting the information I'm posting here from IE's blog..here's the link.

http://blog.internetworkexpert.com/2008/09/24/mstp-tutorial-part-ii-outside-a-region/


Here's my question. I know that MST regions will form a CIST, with their MSTI0's..but does this mean that in the entire STP topology, one switch will function as the root for all MSTI0's..regardless of which region they originate from? I believe that's what the article is saying, and if I recall correctly that's what my lab results were, but I want to double check. Unfortunately I have to rely on rack rentals, so I've got to wait till my time comes up again to verify.


That being said, I have another question. Further on in the article, it discusses mapping MSTI's to the CIST on the boundary. Does this mean if I have say, MSTI5, covering vlans 50-100, that it will show up in MSTI0 (the CIST) in another region? I labbed up a few MST regions the other day, and this didn't appear to be the case..so I want to clarify, because I think maybe I'm just missing something here.

I probably need to review the article 500 times for it to sink in, but I'm hoping someone can at least help me get a grip on this for now.

I appreciate it guys.

Mrock4

OK, I forced myself to re-read the material for a third time..this time without interruptions, and I think I got a good handle on it, so let me pass along what I've learned. Also, I HIGHLY recommend these two articles for anyone studying spanning tree..I love IE's CCIE blog..

MSTP Tutorial Part I: http://blog.internetworkexpert.com/2008/07/27/mstp-tutorial-part-i-inside-a-region/

MSTP Tutorial Part II:
http://blog.internetworkexpert.com/2008/09/24/mstp-tutorial-part-ii-outside-a-region/

Answering my own question regarding the root of MSTIO's...


MST essentially runs two "levels" of STP. The first, IST (MSTI0), which has separate instances if configured, and has a local root bridge, called the IST root, or CIST regional root, which functions as the MSTI0 root for that region only. In a nutshell, MSTI0 can be thought of as having "signaling" information being sent on it, since when crossing boundaries into other regions, the separate MSTI instances will be bundled into the MSTI0 BPDU's...

**
If you're familiar with IS-IS, think of MST like the Level 1 and Level 2 hierarchy. Think of the IST (the local spanning tree MST0) as Level-1, and the CIST between regions as Level 2, with the CIST regional root being L1/L2. The Level 2, in this case the CIST, will not know about any internal STP information, with the exception being the L1/L2..or the CIST regional root, which will act sort of as the gateway to other regions.
**



Moving on, you have the CIST Regional root, which is the root of only that region, and then you have a CIST root..which is the root for all regions. This switch does not see any internal information in regards to MSTI's..it sees the other regions as pseudo-bridges, and acts accordingly. Whatever region the CIST Root falls in, it will also act as the CIST regional root for that region.


As far as interacting with other regions, the mappings will remain the same, however when running MSTP over a link with RSTP on the other side, RSTP will not see the regions. It will see the CST, which essentially looks like another pseudo-bridge to the RSTP process.

Also as a side note, to anyone wanting to configure MST + RSTP on a boundary, bear this in mind. The CIST Root MUST have a lower BID then any switch in the RSTP "domain"..this means the RSTP will see the root as being inside the MST "domain". If a RSTP switch acts as the root, it will do so for ALL VLAN's...meaning your ability to load-balance is completely blown out of the water.


Hope someone else finds some help from this. I'm feeling much better on this. I can't wait till my lab time comes up so I can put it all into action.