1 00:00:00,870 --> 00:00:06,780 This document also provides a lot of information about multiple spanning Tree Hill 82.1 se. 2 00:00:07,110 --> 00:00:12,330 You're not expected to know all the detail in this document, but it provides a nice reference. 3 00:00:12,450 --> 00:00:17,700 If you're interested, I'll cover some of the basics now and then you can read the document if you're 4 00:00:17,700 --> 00:00:19,380 interested in more information. 5 00:00:20,040 --> 00:00:27,240 Multiple spending tree is the new industry standard inspired by Cisco's proprietary, multiple instance 6 00:00:27,240 --> 00:00:28,560 spanning tree protocol. 7 00:00:28,860 --> 00:00:35,640 Cisco developed multiple instance spanning tree protocol to solve some of the issues that you have with 8 00:00:35,640 --> 00:00:36,730 PV ist. 9 00:00:36,900 --> 00:00:44,850 So as the number of VLANs configured in switched networks increases the overhead when running PV ist 10 00:00:44,880 --> 00:00:46,260 also increases. 11 00:00:46,530 --> 00:00:55,230 If you configure a 1000 VLANs with PV ist and rapid PV ist, you end up having a thousand spanning tree 12 00:00:55,230 --> 00:01:00,900 instances, but with multiple spanning tree and the proprietary multiple instance spanning tree that 13 00:01:00,900 --> 00:01:08,880 existed before multiple spanning tree, you can map a number of VLANs to the same spanning tree instance. 14 00:01:09,270 --> 00:01:16,830 It's fairly simple to do this, but the idea is if you had 1000 VLANs, you would allocate 500 to 1 15 00:01:16,830 --> 00:01:23,970 instance and the other 500 to another instance, which means you only have two spanning tree instances 16 00:01:23,970 --> 00:01:26,820 rather than a thousand spanning tree instances. 17 00:01:27,240 --> 00:01:33,930 So multiple spanning tree standardizes, the concept of multiple spanning trees and incorporates the 18 00:01:33,930 --> 00:01:36,630 convergence of rapid spanning tree. 19 00:01:36,780 --> 00:01:43,020 Multiple spanning tree allows you to group VLANs to a shared spanning tree instance. 20 00:01:43,260 --> 00:01:47,710 It also defines a protocol for interconnecting multiple spanning tree regions. 21 00:01:47,730 --> 00:01:54,840 How to interoperate with existing ADA 2.1 DX and ADA 2.1 CU spanning tree implementations and provide 22 00:01:54,840 --> 00:01:56,190 some best practices. 23 00:01:56,400 --> 00:01:59,460 So have a look at this document if you're interested in a lot of detail. 24 00:01:59,820 --> 00:02:01,530 But here's a quick comparison. 25 00:02:01,920 --> 00:02:09,539 Imagine you had 1000 VLANs on switch A which is connected to both switch D one and DH two. 26 00:02:10,199 --> 00:02:17,640 So in this topology switch A has a 1000 vlans d one is going to be the spanning tree route for some 27 00:02:17,640 --> 00:02:22,170 VLANs and DX two is going to be the spanning tree route for other VLANs. 28 00:02:23,640 --> 00:02:31,100 So switch D1 is configured to be the route for VLANs 501 to 1000 DH two is the route for VLANs 1 to 29 00:02:31,100 --> 00:02:31,530 500. 30 00:02:32,070 --> 00:02:39,330 The interface from switch A to switch D one blocks, VLANs 1500 and from switch A to DH two blocks, 31 00:02:39,330 --> 00:02:41,730 VLANs 501 to 1000. 32 00:02:42,270 --> 00:02:51,060 So once again, route for these VLANs traffic will be forwarded out of this port for those VLANs but 33 00:02:51,060 --> 00:02:53,760 blocked for VLANs 1 to 500. 34 00:02:54,270 --> 00:02:57,090 The switch is the route for these VLANs. 35 00:02:57,810 --> 00:03:06,450 This port will forward traffic out of it for VLANs 1 to 500 but will block a VLANs 501 to 1000. 36 00:03:06,780 --> 00:03:13,200 It's very inefficient to maintain a thousand spanning tree instances in this network. 37 00:03:13,980 --> 00:03:23,370 We have 500 spanning three instances with D, one is the root and we have 500 with RD two as the root. 38 00:03:23,550 --> 00:03:27,690 But logically we actually only require two instances. 39 00:03:28,140 --> 00:03:35,790 RD one should be the root, for instance, one that contains these VLANs and RD two should be the root, 40 00:03:35,790 --> 00:03:39,480 for instance, two that contains these VLANs. 41 00:03:39,930 --> 00:03:47,820 You associate these VLANs to instance one and make rd one the route you associate these VLANs to insulins 42 00:03:47,820 --> 00:03:48,420 to. 43 00:03:49,070 --> 00:03:50,870 And make D2 the route. 44 00:03:51,410 --> 00:03:55,910 That means you have to maintain two instances rather than a thousand instances. 45 00:03:56,970 --> 00:03:59,430 So that kind of detail is explained here. 46 00:03:59,460 --> 00:04:00,750 I'll go through it quickly. 47 00:04:01,140 --> 00:04:06,870 In a Cisco PvZ environment, you need one spanning tree instance for every VLAN, which means you have 48 00:04:06,870 --> 00:04:13,770 a 1000 instances for the two different final logical topologies with DX. 49 00:04:13,770 --> 00:04:17,880 One is the route for one topology and DX to the route for the other topology. 50 00:04:18,000 --> 00:04:23,220 This wastes a lot of CPU cycles for all the switches in the network. 51 00:04:23,550 --> 00:04:30,930 In addition to the bandwidth used by sending BPD use, a thousand BPD use are going to be sent out of 52 00:04:30,930 --> 00:04:40,260 every port every 2 seconds because PVS RT sends a BPD queue for every VLAN because we have an individual 53 00:04:40,260 --> 00:04:42,390 instance mapped to every VLAN. 54 00:04:43,600 --> 00:04:51,220 So the idea with multiple spending tree is that you get the best of PVS tree and traditional spending 55 00:04:51,250 --> 00:04:51,820 tree. 56 00:04:52,030 --> 00:04:55,780 You map several VLANs to specific instances. 57 00:04:55,780 --> 00:04:57,790 So in our topology once again. 58 00:04:58,480 --> 00:05:06,100 You'd make switch one the route, for instance, one switch to the root, for instance, to the sport 59 00:05:06,100 --> 00:05:11,080 would forward, for instance one but block, for instance, to the port would forward, for instance, 60 00:05:11,080 --> 00:05:19,540 to but block for instance, one only two spanning trees are maintained rather than a thousand. 61 00:05:20,320 --> 00:05:26,350 So you still get load balancing because half of the VLANs follow a separate instance and you save on 62 00:05:26,350 --> 00:05:29,710 the CPU because you only have two instances of spanning tree. 63 00:05:30,010 --> 00:05:36,100 So from a technical point of view, multiple spanning trees, the best protocol to use in this example, 64 00:05:36,100 --> 00:05:44,050 but multiple spanning tree is more complex to configure than PV ist and interaction with legacy switches 65 00:05:44,050 --> 00:05:46,090 can be challenging at times. 66 00:05:46,630 --> 00:05:52,050 So you would only want to use multiple spanning tree if you have many VLANs. 67 00:05:52,060 --> 00:05:57,040 So in this example we have 1000, so it makes sense to use multiple spanning tree. 68 00:05:57,160 --> 00:06:06,430 If you only had ten or 20 VLANs in your network, you could continue using PV RT or rapid PV IST. 69 00:06:07,600 --> 00:06:13,810 The document continues with a lot of detail on how to configure multiple spanning tree regions, but 70 00:06:13,810 --> 00:06:16,030 that is out of the scope of the Q&A. 71 00:06:16,240 --> 00:06:18,790 So have a look at this document if you're interested. 72 00:06:18,880 --> 00:06:22,510 We'll wait till you get to your CMP certification.