145 IPExpert Multiple Spanning Tree MST

[Music] perceived as the most complex of the tree spanning tree mode supported by Cisco switches multiple spanning tree or the MST is going to be our next topic I'm going to do now something that I haven't done a lot in this video series but I'm going to do it now please before you proceed and watch this video about the MST I beg you watch the other videos that cover other operational modes of spanning tree the prevalence spanning tree and the rapid prevalence spanning tree the reason for that is that MST is perceived complexity comes from the fact that it draws on all those concepts of other spanning trees and then it builds on top of that so if you were struggling with concepts from especially rapid prevalence penalty or the rapid spanning tree you are going to have some difficulty understanding what MST does so what I'm going to do now is I'm going to ask you to go back watch those videos again get those concepts solidly down so that we can proceed and explore what is actually different in MST in comparison to those other modes of operation of spanning tree on Cisco switches of course if you have already done that and just recently watched those videos now is the time to actually proceed the first thing to note about the multiple spanning tree or the MST is that it's fundamentally based on the rapid spanning tree and all the good stuff that comes with rapid spanning tree things like edge and on edge ports they are still there all the handshakes in synchronizations all those processes are there for the MST so all the good stuff all the benefits from the rapid spanning are there in the MST now MST operates in something that we call the MST region that's a group of switches that have the same that share the same MST configuration the MST configuration consists of the MST region name which is basically a string up to I believe 32 characters long then there is a vision number which is just an administrative number that you said so unlike VTP where it's increased with any addition or removal of the villains or any configuration with the MSD configuration against the revision number this is an administrative number it's a number that you simply set yourself finally there is a villain to instance map now remember MSD supports multiple instances I'm going to talk about that in great detail in just a moment but remember that what defines a region are these three things pick a mysterion name the revision number and the villain to instance mapping now the real complicated stuff really complicated complex part of operation of the MST comes when we are talking about the backwards compatibility of the MST with other types of spanning tree with other types of spanning tree known cisco implementations of spanning tree like for example 8o 2.1 d or e 2.1 q as well as cisco proprietary implementations of spanning tree like p vs t and the rapid p vs t to operate in a backwards compatible mode MST divides the whole network that means MST region other MST regions and non MST switches into basically several parts of that network now that whole entire network is something that the MST calls the common spanning tree all the CST inside the MST region there is an operation of the spanning tree which is invisible to the outside world the world outside of that particular MST region which is called the internal spanning tree or the ISD technically speaking this is instance number 0 but I'm going to get back to that in again just a moment and finally inside that same region all different instances and remember villain 2 instance map is what defines these instances are operating in something she called the MST I or the MST instance spanning-tree there is a lot of stuff here that may sound confusing and maybe a lot to fathom so let me go I had an explained couple of these concepts so let's talk about the MST regions as I mentioned before the MST region is a group of such as a network of switches in the same network interconnected together the chaired the same configuration of the MST region name or sometimes called domain name which is a string of characters up to 32 characters long and then we have the villain two instance mapping now the important thing to understand here is that MST BPD use that I'm going to talk about in a moment about they don't actually carry this villain to instance mapping what we are going to observe in the BPD use is going to be a hash value that is going to be computed based on the local value of instance to villain mapping or villain to instance mapping now when VP user received these hash values are compared and switches know whether this value is the same or is not the same and finally we have the configuration religion and as I mentioned the configuration revision is just a number and administrative lease at number that you can configure to be any number in the range from 0 to 60 to 65535 now as you can see here the configuration of these parameters couldn't be easier all you have to do is say spanning tree MST configuration you specify the main revision and instance mapping or v12 instance mapping and you are done let me talk about default values for these just a little just for a little bit the default value for the main is simply an empty string there is not going to be a name but if there is going to be a value an empty string which is still a value so technically speaking where I said there that the name is up to 32 characters long I should have mentioned that it could be 0 characters long as well and it's still a valid the name the next is the revision number the default revision if you don't set one manually is going to be zero and instance mapping or the villain to instance mapping by default all villains are going to be mapped to instance zero and in Cisco IOS you can have up to sixteen instances defined or to put this in more practical terms let's say that we have a couple of switches in our network and these switches here are configured with these same parameters and literally they are interconnected doesn't matter how they are actually interconnected so let's say these switches here are sharing the same domain name they are sharing the same revision number and they are sharing the same instance to VLAN mapping or villain to instance mapping so I'm just going to call this a map so let's say these switches here they belong to the same MST region so if these three here are the same we can say that these switches here belong to the same MST region but let's say that we had another set of switches and these which is here may have had a different name so let's say that here we had some name number two we had some different sort of revision and we had a different or the same map but at least the name is different so these value he values here are the same and these switches here would be in a different MST region so this interconnection between these which is here this is not going to be the same region the different rules are going to be in action there and those different rules I'm going to cover when I talk about the backwards compatibility with other spanning trees now in this particular sense here this is not a backwards compatibility with a different kind of spanning tree this is the simply compatibility with the MST that runs in a different region so inside this region here and inside this region here one set of rules applies and here between these two regions are different the rules apply what I mean by set of rules is what kind of spanning tree we are running there and remember when it comes to the MST there are different kinds of spanning tree we have the internal spanning tree we have the common spanning tree and we also have the MST instance spanning tree running so with all this in mind let's take a look at how do we actually configure the MST and how does all this information look when it's transmitted in BPD use to do that I'm going to use a very very simple Network consisting of just two switches we don't need more than two switches to start with so what I'm going to do here is I'm going to use my trusty cat 1 and cat too and I'm going to interconnect them with fast new 0 24 on both sides and on cat 2 for example I'm going to connect my Wireshark listening station on Fast Ethernet 0 3 and just as I have been doing so far I'm just going to send the copy of this traffic receive them sent on fast in at 24 to my Wireshark here so that I can examine what is happening between cat 1 and cat too but before I do that what I'm going to do is I'm just simply going to send enable this port and see what kind of BP news I am receiving without actually observing what is happening between cat 1 and cat 2 and also I'm going to configure these tools which is here to be in the MST region that is going to have the name let's say Cisco we are we are going to have revision let's say revision 1 and I'm going to leave the default map so that all villains are going to be mapped to instance number 0 so let's see how this configuration is going to look so I'm going to go to my terminal now and here is my cat 1 and I'm simply going to say spanning tree MST configuration and here I have several option to do now keep in mind one thing this is very very similar to our villain database it's not the same but it's very similar it's very similar in a fashion that whatever I configure here now will not take immediate effect I would have to exit out of this configuration mode for the changes to actually take effect let me show you what I mean by that if I do show spanning-tree MST configuration this is the default configuration that I have right now so if I do name and I set the name to Cisco if I repeat my show MST configuration take a look at what it says here it says first of all that which is not in the MST mode which is okay but the name is still unchanged if I repeat the command now I will see that the name has been changed to Cisco so we can see here that the name changed to Cisco only by virtue of me exiting the configuration this also means that in the case I have changed my mind so let's say that I went into my spanning tree MST configuration and I set revision number 210 if I exit this is going to take effect but I can use the command abort and with the abort command I'm not going to have any changes affected here so I can see that the revision number stays zero now also another thing that I would like to show you is that by default we do have an instance on figure we can see here that we have one instance configured and this is going to be our instance number zero and we can see that all villains by default are mapped to this one instance so let's go ahead and change this configuration to be what we actually want it to be so I'm going to say spanning tree MST configuration here I'm going to say named Cisco I'm going to say revision number one and if I do show spanning tree MST configuration this is what I will be seeing remember until I exit the changes I have made do not take effect but now I can see that the change actually was affected and I can see that revisions has been changed to revision number one this is something that you may want to keep in mind when you are configuring the MSD because it's very very easy to get carried away and think that your configuration actually was affected was executed has been committed to the running configuration but it wasn't until you actually exit from this sub configuration mode but this said we still have to keep in mind that switch even though we have MSD configuration is not actually operating in the MSD mode to do that we actually have to change the operating mode of the spanning tree on our switch we can do that with spanning tree mode and we can see here that we have options MST previously which is the default and rapid previously so what we want to do now is we want to change the operating mode to be the MSD with this change in place now our switch operates in the MSD mode we can confirm that if you do show spanning tree what we are going to see here is that spanning tree is enabled and that the protocol we are running is MSTP and if we just to show spanning tree what we are going to see is going to be the operating modes for all the configured instances and by default we have only instance number 0 active and we can see here that for example portfast in a 24 is now in a learning phase if you watched rapid a pv st videos you will know that this is not supposed to happen if both sides are operating in rapid mode which is one of the prerequisites for the MSD this should have already transitioned into forwarding using the synchronization process and the handshake process but it didn't because this port here is actually a boundary port so one thing that I want to point out about this name here it doesn't mean that this port is bound to anything this is short of a boundary this is telling us that this port fasted in a 24 is connected to a different other MST region or a different spanning tree domain what I mean by spanning tree domain is that this other side is running as far as this which is concerned regular spanning tree now we know it's running previously because that's the default but this is how we are going to see it if we do show spanning tree we are going to see that this is a boundary port and then that on this port we actually have to fall back and operate in traditional spanning using traditional spanning tree rules just like the rapid spanning tree or rapid prevalent spanning tree would do it would fall back to normal traditional operations of the spanning tree if it detects different kind of BP news coming from the other side let's then correct this well technically speaking there is nothing to correct but let's finish our configuration properly let's go ahead and configure cat to to operate in MST mode and I'm going to configure it using the same parameters that I configured on the other side so what I'm going to do here is I'm going to say a spanning tree MST configuration I'm going to say that name is Cisco revision number is 1 and I'm going to see I'm going to leave my villain 2 instance mapping intact I'm just going to exit from this configuration road and if I do show spanning tree MST configuration this is what I should be seeing now let's go back here and say a spanning tree mode MST so now I've changed my cat to to operate in the MST mode and if I go back to cat 1 what I should be seeing here is that this is no longer a boundary port so this is no longer effective this is now the port and I can see that this is now just a regular point-to-point port because the other side is operating in the MST mode so this was very very basic configuration of the MST this was very very simple now let's see those BPD O's let's see how these BPD use look on the wire and what is the information carried in the configuration be produced for the MST to do that obviously I'm going to go to my Wireshark and as I said this Wireshark has been connected to port fast it'll at 0 3 I'm not going to configure any mirroring I'm just going to enable this port on my cat too so if I take a look at the configuration of faster than 0-3 I'd see that the port is shut down and I'm just going to bring it out of shutdown and going to my Wireshark I'm going to make sure that I'm capturing only the spanning tree BPD use and I'm going to start the capture here so let's take a look and how these BPD look so I can already see that I'm receiving some traffic so I can stop this capture I don't want it to scroll up and down and let's examine what are the contents there the first thing that I would like to point out is that we can see that the protocol version identifier is MST it's going to be 3 so we see now here that this is a different kind of the BPD oh we can see that this is a BPO type that is actually either rapid or the multiple spanning tree and also if I scroll down here there are going to be some extensions now these MST extensions this part here this part of BPD is what we are interested in looking this is where the information about our configuration is going to be transmitted and exchanged between the switches the first thing that we can see here is the MST configuration name and we can see it is what we configured it to be it is set to Cisco we can see the configuration revision 1 but this bit here is the interesting bit this here is villain 2 instance mapping and as you can see it's fairly unreadable by humans this is just a really really large digest now with this digest here the switches will know if the configuration is the same and it's important to understand that this digest will change when the villain to instance mapping has been changed let me demonstrate that so I'm going to restart the capture so let me start the capture again and I'm going to be getting some reproduce what I would like you to focus your attention is this field here so we can see that these this value here remains the same some now moving to different videos that are being transmitted and we can see that this value here the VIP configuration digest does not change so let's go back to terminal and just for fun I'm going to say spanning-tree MST configuration and I'm going to change the revision to 3 remember it was one just before so going back here at this point here if I scroll down let's see where the configuration change happened so here at this point we can see that the configuration revision is 3 and this is the digest value that we are looking so I'm just going to scroll one packet up and we can see that the configuration revision is 1 but the actual configuration digest is unmodified so changing the revision does not modify this value here ok fair enough let's change the revision back to what it was and let's change the name to something else let's change it to Cisco so I'm going to go back and apply this and let's scroll down where this value has been changed so there it is now we can see that the domain name Cisco all uppercase is there but this value here is unchanged in comparison to what it was before so we can see here that this is where the change happened and we can also see that a revision number here is going to be changing so this is where it was 3 and this is where it changed to 1 and all the uppercase but we can see that the configuration digest remains the same so this configuration digest there that field is where the information about villain to instance mapping is carried let me show you that as well so going back to the terminal on count two and we'll say spanning tree MST configuration and I'm going to say instance so now I'm going to create a new instance and I'm just going to move villain number 10 there please take note villain number 10 does not exist on the switch at this moment so if I do show villain I'm not going to see wheeler number 10 and this is not actually relevant because this is a role-playing information this is just information what are we going to do what kind of logical topology we are going to be building not exactly what villains we have so spanning-tree does not carry the information or the MSD does not carry information only about the active villains because we didn't have 4,000 94 villains to begin with but it carries information in a case we have that traffic in a case we want to have traffic in that wheel and this is the logical topology the death traffic would be following so unlike per villain spanning-tree where we have one instance for every created villain here we may have villains map two instances and those villains may not be in use or not all of them could be in use and this is perfectly normal because that's the whole idea behind the MST is that we are not going to have a large number of instances they're going to have a limited number of instances and I'll explain why in a moment why this is actually good idea to have so but let's see what this little change actually did for our VP views so at this moment this change has not been applied to fellow show spanning tree MST configuration I'm going to see that at this moment all villains are still mapped to instance 0 going back to up to my Wireshark let me scroll all the way down here because this is where I want to start looking at it so now when I exit if I repeat my show spanning tree MST configuration now I'm going to see that the instance that the villain 2 instance map has indeed been modified what I can see here is that for instance 0 these villains have been mapped and for instance one only villain 10 has been mapped to it and if I go back to my Wireshark here where was that boundary packet there it is so let's see where this thing actually changes should have been changed somewhere around here oh did I actually apply the changes as I have there it is this is where the boundary happened so at this point here so this is the the frame where we can observe the difference so at this point here please take note what the MST configuration digest is now the next bpdu here carried a completely different digest it carried the different digest because our villain to instance map has changed but the actual map is not transmitted in the be produced it's locally kept on the switches and then based on that configuration alone the MST configuration digest is computed so this tells us how to configure it but it doesn't tell us why are we configuring these multiple instances well let me explain that a little bit let's say that we had again just two switches and let's say that for the sake of this example these two switches were interconnected by two links and let's say that in this network we had 10 villains for the traditional implementations of the spanning tree in Cisco switches which done by traditional here I really mean PDS team or rapid previously what we are going to have here are going to be ten different logical topologies by ten different logical topologies what I really mean is ten spanning tree instances now with these ten spanning tree instances what I have are really just two possible logical loop free topologies given my physical diagram so let's say this is which one and this is which two that this is link one and this is linked to now one of the possible logical topology that we can have for this physical diagram here is that switch one and switch to have link one forwarding and that link two is blocking on either side that that doesn't matter another logical topology so this is logical topology one and another logical topology here is that we have switch one switch to where link two is forwarding and then let's say that link one is booking so we have two possible logical topologies but regardless of that fact we are still going to be using ten different instances of spanning tree now if there is an overkill this is one so in order to address this the MST actually allows us to map multiple VLANs to a single instance of spanning tree so to solve this problem to utilize all links in this network topology here we don't need ten instances of spanning tree we just need two instances of spanning tree so let's say that in our example here these villains were villains one to ten what we could do is we could map for example Sweden's 1 2 3 4 & 5 - let's say instance 0 and we could map 6 7 8 9 and 10 to instance 1 then what we could do we could make this switch here root for instance 0 and we could make this switch here root for instance 1 for example and based on that we could also tweak the costs on different links here or we could trick the priorities on different links and come up with different topologies that actually result in different links forwarding or we don't even have to make different routes in this case we could make switch 1 be the root for both instances and then tweak the costs on these links here or these links here to be more precise in order to affect which link will be blocking and which link will be forwarding let's see this in action to do that I'm just going to continue with my cat 1 and cat 2 example and I'm just going to replicate whatever I have just described so I'm going to have my cat one and cat - I'm going to have two links between them this is going to be linked fast in a 24 and this is going to be linked fast in at 23 and I'm going to create villains from 1 to 10 let's start by doing this so I'm going to go to terminal here and I'm just going to say villain 1 to 10 going to create those villains let me do that on both switches so if I do show villain here I'm going to see that I do have now my ten villains created on switch 1 and I can on my cat 1 and I can see that I have those ten villains created on cat - the next thing that I'm going to do is I'm going to go to spanning tree MST configuration and I'm going to say instance 0 villains from 1 to 5 and from 11 to 4 thousand 94 so I want to have the default specific all the other villains I'm going to keep them in instance 0 and this thing here is one more thing that I think is very neat to know so here I can use the same specification that I would be using what I'm defining or allowing villains on a particular portrait which ports trunk allowed villain so I can create either the range or I can list the villain so for example here I have two ranges from 1 to 5 and from 11 to 4 thousand 94 and they are separated by a comma so I can specify Leon's in any sort of way that I that I really really want now let's see this configuration so if I do show spanning tree MST configuration here I can see it okay well I have not mapped those other villains too so I have created the the map for instance 0 but what I need to do is I need to move those other villains to instance 1 1 so here I'm going to say from 6 to 10 it's actually going to go to instance 1 so if I do show spanning tree MST configuration now I see that villains from 1 to 5 and 11 to 4000 94 are mapped to instant zero and six to 10 are mapped to instance one let's do the same thing on cat too so I'm going to say instance zero villain from 1 to 5 and from from 11 to four thousand ninety four and instance one villain six to ten show show spanning tree MST configuration I can see that my configurations on cat 1 and cat two are the same well almost the same let's not forget that the the domain should really match let's confirm that the configurations are now the same it's easier to do it side by side so I'm just going to do it this way and I can see that on switch one and switch to the configurations are the same going to cut one if I do show spanning tree MST now I'm going to see that only MST 0 appears to be active what happens on cat to show spanning tree MST it's the same thing why is it showing me only instance 0 as active when I have created all the villains well the reason here is that there are no ports that utilize any villains from the range that is mapped to instance one so that instance one is not activated similarly to what you would have with the rapid preview on spanning tree or the normal prevalence spanning tree or traditional prevail on spanning tree is that if you don't have any ports that require the instance of the spanning tree it's simply not going to be activated so how do we add ports to any of those villains well very simple what I'm going to do is I'm going to configure the range of ports or interface range fast in the 23 to 24 I'm going to declare them strong so switch port trunk encapsulation dot1q switch port mode drunk and I'm going to do now shut down here interface range faster than 23 to 24 switch port ranking capsulation dot1q and switch port mode trunk no shutdown so now at this moment if I do show spanning 3 MST what I should be seeing is two instances of spanning tree running and that's exactly what I can observe now I have MS teen 0 and I have MST one now another observation that we have to make is that this switch here is this switch is the route switch for the internal spanning tree and this one here is the root for MST instance one I wanted to reverse this I wanted to use both ports and if I have this set up like this I'm going to have 23 and 24 designated ports on both sides sorry for both instances of spinelli but this is not what I want on the other side on the other side what I want is actually slightly different configuration what I would like to have is that for one instance 24 is blocking and that for other instance it's actually 40 so that I can utilize both lengths that is easily easily done so this is what I'm going to do on cap - I'm going to go to interface first in the 23 I don't want to say spanning tree MST instance 1 for example because I'm going to increase this cost let's say that I'm going to set the cost be four thousand three 450,000 so if I do show spanning tree MST now what I'm seeing at this point here is that for instance 0 port 23 is forwarding and for instance 0 for 24 is blocking but for instance 1 the roles are actually reversed I can see here that port 23 4 villains to them is going to be blocking but it's going to be 44 villains one to five now this is the reason why MST exists to begin with you see now that Ford the traditional implementation of per villain spanning tree or the rapid per villain spanning tree that cisco implements on catalyst switches we would have needed a lot more instances to actually accomplish this with the multiple spanning tree we just did two instances and we are utilizing both links this is it as simple as it gets multiple spanning tree MST on its own is rather complex spanning tree operation however the most intimidating and most complex part of it is how it actually interoperates and cooperates with different kinds of spanning trees so in this module we are going to talk about MST compatibility with the rapid spanning tree rapid pdsd we are going to talk about MST is interoperability with the Cisco's PV st and of course what happens when we have multiple MST regions in a single Network we can run different kinds of spanning tree we can run MST the PDS the rapid PDS T we can run traditional spanning tree if you're running non Cisco switches we are using on Cisco switches and all these different parts of our network that are running different kinds of spanning tree are considered to be spanning tree domains so a simple spanning tree domain would be a single switch where all ports are in VLAN 1 and we have a single instance of prevalence spanning tree that one switch makes up a single spanning tree domain on the other hand we could have a more complex spanning tree domain that consists of an MST region consisting of let's say 10 switches and five instances inside that MST that has 150 VLANs mapped across these different distances this is also considered to be a single spanning tree domain this entire network can still work and interoperate and can function I'll bet it might be a little bit difficult to troubleshoot for any inexperienced or even more experienced network engineer but it will still work somehow so to see how that kind of network will work let's take a look at our spanning tree domains so let's say that we had a part of our network that was running traditional spanning tree so let's say the D switches here are running traditional spanning tree let's just make them Cisco switches so these are running p vs d spanning tree let's say that we have couple of switches there that are running that are running the rapid prevalent spanning tree and they're somehow interconnected here and let's say that we have an MST region where we have different switches that are running the MST so these regions here these enmity domain the spanning tree domains could be interconnected and we don't really care how the internal connections look here because this is all going to work because both rapid prevalent spanning tree and the MST they are backwards compatible with the PDS T so this is somehow going to work out however MST takes this compatibility with different kinds of spanning tree one step beyond this simple backwards compatibility this operating in the old mode as we know all these ports that are connecting to different kinds of spanning trees we are going to call them boundary ports and Cisco is simply going to display them as bound ports remember these are not bound to anything these are bound the reports these are the border ports in this which is here we can call them the border switches because they are sitting on a border of our MST domain and something else what we can say from MST is perspective so we are looking at the MST is perspective of this network we can say that this whole network here is something that we call a common spanning tree so this is a global view of the network this is the spanning tree that keeps this whole network loop free however inside the MST we still want to preserve our visibility into different instances into different topologies that we might have but we really want to hide all that complexity from the outside domains that don't actually understand what MST does so from for that reason inside our MST domain inside our MST region we are going to be running something that is called the internal spanning tree now internal spanning tree is nothing other than the MST instance 0 but what happens if we have multiple instances well if we have multiple instances these same BPD users are going to be used to carry this information but they're going to have information about something else the instance the MST instance spanning tree there so we are going to have information about let's an instance one or instance to our instance tree and so on so we can say that in our spanning tree network in this large network that consists of multiple spanning tree domains we are really running from MST is perspective three kinds of spanning trees we are running the common spanning tree we are running the internal spanning tree and we are running for instance spanning trees inside the MST domain if you are asking yourself what's the reason for this complexity that we are seeing here I don't blame you I'm struggling with the same concept for a very very long time and then it struck me even though this whole network would operate without this complexity deterministic introduces it actually makes Network run just a little bit better just a little bit smoother than it would without this complexity and that's the whole goal in the networks we want the networks to run in a more optimum way we don't want to have network that is backwards compatible and slow when something fails we know that T vs T will take a long time to converge after the failure rapid PVS tea is somewhat better but if we all switch back to our backwards compatible operational modes what we are going to end up doing is running PVS tea everywhere or traditional spanning tree which is not the champion of speed when it comes to convergence so for that reason MST introduces this additional level of complexity to make things like run just a little bit faster and a little bit better when things go wrong in our networks and as we know from time to time things do go wrong the basic idea is relatively simple it is let's hide the MST region from everyone else by representing it or presenting it to everyone else as a single switch but inside the MST region let's maintain the knowledge of what is happening inside the region itself in other words from the perspective of the MSD the network may look like this so let's say that we have three switches that are in the MST region and let's say that we had a single switch that runs PDS T here so this may be the actual network or how the actual network looks from the perspective of the PDS T what BST is going to be seen here is going to be the PDS tail switch connected to another PDS T switch with two links now we know that this switch here is actually our whole MST region and in our MST region we are running our internal spanning tree with probably it installed the internal spanning tree instances or MST eyes actually and here we are running the common spanning tree so here we have our internal spanning tree that maintains this topology and then here using our common spanning tree the PDS T is going to perceive this whole MST region this whole set of switches the group of switches as a single switch this allows for more optimum operation inside the MSD domain while maintaining relatively simple perspective for everyone else

Channel: CCIEORDIE.COM

Views: 8,361

Keywords:

Id: NeULp-FXAfY

Length: 44min 41sec (2681 seconds)

Published: Thu Feb 08 2018