BCMSN Question - MST?

Ok, from what I understand from just finishing chapter 5 in the study guide, MST is a rapid spanning-tree mode that can help reduce the number of spanning-tree instances that a switch is running if it has a ton of vlans, however, it requires a bit more complex configuration as well as consuming more switch resources...so...my question is, why would you want to do this then? Is there more rapid convergence after a topology change? Also, how do you know when you have a number of vlans large enough to justify the additional configuration and resource overhead of MST?

I is confused.

Find more posts tagged with

Free for TechExams community: Cybersecurity salary guide

Compare cert salaries and plan your next career move

Button

Comments

kryolla

There is a spanning tree instance for every vlan created. If you have multiple vlans that traversing the same path why have all these instances and overhead? Just setup MST and put all these vlans in the same instance. It depends on how much control you want per vlan for traffic engineering. If you have an access layer switch with 2 uplinks how much traffic engineering can you do so if you have 50 vlans that is a lot of overhead just make 2 instances. I haven't worked too much with MST so take this with a grain of salt.

Mrock4

Building off of what kryolla said, MST will allow you to dramatically reduce the amount of STP instances, but still retain the ability to load balance between them. If you had say, VLANS 1-10, you could tell Vlans 1-5 to take path A (note- in only one instance, not multiple), then 6-10 to take path B in a second instance. Instead of 10 instances, you now have two. You can do this on a much larger scale obviously..

Netwurk

I had MST configured in my lab and it worked great

Posted my config about a year ago in the http://www.techexams.net/forums/ccnp/29382-finally-configured-mst.html thread.

But to get everything to sink in (for the BCMSN), I've now moved my lab back to PVST. From there, I'm going to migrate to RSTP and then finally back to MST.

To me, MST is easier to manage than the other types once you do all the work to get it up and running. You get to group your VLANs with instances, rather than deal with something like the 50 VLAN scenario that Kryolla outlined.

You can still have 50 VLANs of course, but you can group them more easily.

redwarrior

I just started a project to try this out at work.

At every building on each campus, we have 2 4500 or 4900 series switches with multiple redundant links. Right now, only 1 of them is acting as root for every vlan they share. I'm going to lab it out and propose a project to convert them to MST with half the vlans rooted on one switch and half on the other. It seems silly to waste those links when they could be used for load balancing.

unclerico

This may or may not be an issue, but make sure your switches are running at least 12.2(25) code in order to use the IEEE 802.1s version of MST, not Cisco's pre-standard MST. This will be very important if you ever need to integrate switches from other vendors into your spanning-tree toplogy as they will support the IEEE version. If you run pre-standard MST on your Cisco's, but IEEE standard on your other switches then no matter if you have the config name, the revision number, and the vlan mappings the same they will never be part of the same region (the ports they are connected to will be flagged as boundary ports). With that being said, if you are running all Cisco everywhere, you can run pre-standard and standard and they will function ok.

Forsaken_GA

unclerico wrote: »

This may or may not be an issue, but make sure your switches are running at least 12.2(25) code in order to use the IEEE 802.1s version of MST, not Cisco's pre-standard MST. This will be very important if you ever need to integrate switches from other vendors into your spanning-tree toplogy as they will support the IEEE version.

Yeah, I can confirm that this particular issue is something to consider if you plan on trying to make force10 and cisco gear play nice. Preferably before you bring your entire network down.