1 00:00:00,300 --> 00:00:07,350 In a topology like this, we had to configure spanning tree, HSRP link aggregation and other technologies 2 00:00:07,350 --> 00:00:11,220 to try and optimize the way that the topology works. 3 00:00:11,880 --> 00:00:18,990 As an example, we made a switch 1 route for some VLANs, and then we made a switch 2 route 4 00:00:18,990 --> 00:00:20,210 for other VLANs. 5 00:00:20,610 --> 00:00:29,670 So we managed to optimize Layer 2 forwarding by optimizing spanning tree, but then to optimize the routing 6 00:00:29,910 --> 00:00:38,910 in this topology. We had to configure HSRP and then configure switch 1 as the HSRP primary for the 7 00:00:38,910 --> 00:00:43,470 same VLANs as the as the route switch. 8 00:00:44,120 --> 00:00:49,890 In other words, if switch 1 is the spanning tree route for VLAN 10, then we had to make it to the 9 00:00:49,890 --> 00:00:52,840 HSRP primary router for VLAN 10. 10 00:00:53,220 --> 00:01:00,190 So in other words, we have to optimize both spanning tree as well as HSRP to make sure that they are 11 00:01:00,210 --> 00:01:01,010 aligned. 12 00:01:01,800 --> 00:01:07,770 In other words, if switch 1 is the spanning tree route for VLAN 10, we don't want to make switch 2 13 00:01:07,980 --> 00:01:14,610 the HSRP primary router for VLAN 10. We want to match them up to optimize the forwarding. 14 00:01:15,120 --> 00:01:17,880 So there's a lot of additional work that you need to do here. 15 00:01:18,300 --> 00:01:19,730 You need to configure spanning tree. 16 00:01:19,740 --> 00:01:24,990 You need to configure HSRP, you need to configure link aggregation, you've got to match it all up. 17 00:01:25,140 --> 00:01:26,600 It's not very efficient. 18 00:01:27,060 --> 00:01:29,040 Is there a better way of doing this? 19 00:01:29,880 --> 00:01:34,020 And fortunately the answer is yes, there is a better way of doing this. 20 00:01:34,380 --> 00:01:41,940 I've done a search in Google for Cisco stacking images and you'll find many images such as the following 21 00:01:42,120 --> 00:01:46,110 showing you examples of Cisco switches being stacked. 22 00:01:47,600 --> 00:01:49,970 Now, there are different technologies for doing this. 23 00:01:50,940 --> 00:01:56,310 In other words, there are different stacking technologies, one of the oldest is stackwise, which 24 00:01:56,310 --> 00:01:58,860 was used on the 3750 switches. 25 00:01:59,770 --> 00:02:05,050 But again, if you do a search in Google where you have a look at some of the Cisco documentation, 26 00:02:05,260 --> 00:02:11,870 you'll see a lot of examples of different stacking technologies that can be used to stack Cisco switches. 27 00:02:12,640 --> 00:02:15,540 So what's the advantage of stacking switches in this way? 28 00:02:17,350 --> 00:02:23,020 In brief, when you stack switches, they appear to be a single switch to the rest of the network, 29 00:02:23,620 --> 00:02:25,610 you configure it as a single switch. 30 00:02:26,080 --> 00:02:27,880 They act as a single switch. 31 00:02:28,450 --> 00:02:34,690 Protocols such as spanning tree and CDP, see that switch as a single switch. 32 00:02:35,170 --> 00:02:41,860 So, again, Cisco has various technologies that you can use to stack switches or aggregate Chassis. 33 00:02:42,310 --> 00:02:49,150 So we have these terms, switch stacking and Chassis aggregation, separate physical switches, work 34 00:02:49,330 --> 00:02:56,320 and cooperate together to act and appear as a single switch rather than multiple discrete switches. 35 00:02:56,950 --> 00:03:02,710 As an analogy, it's as if the switches are acting as blades in a chassis-based switch. 36 00:03:03,280 --> 00:03:11,200 Switch stacking is often used at the access layer and Chessie aggregation is often used at the distribution 37 00:03:11,200 --> 00:03:12,980 and core layers of a network. 38 00:03:13,720 --> 00:03:20,410 So if you've got multiple layer access switches or multiple distribution or Core layer switches, rather than 39 00:03:20,410 --> 00:03:27,510 you having to configure each switch individually and configure protocols such as spanning tree, VTP, 40 00:03:27,550 --> 00:03:33,520 CDP and so forth, you configure multiple physical switches as they're one switch. 41 00:03:33,910 --> 00:03:35,380 They have a Mac address table. 42 00:03:35,830 --> 00:03:41,370 They run protocols as if they were a single switch and share a Mac address table. 43 00:03:42,010 --> 00:03:48,070 So when you stack switches at the access layer, you're literally creating a stack of switches joined 44 00:03:48,070 --> 00:03:50,050 together via special cables. 45 00:03:50,800 --> 00:03:57,700 So a stack of physical switches in a wiring closet as an example can act as a single switch. 46 00:03:58,150 --> 00:04:05,140 You would manage the stack with a single management IP address, you would telnet or SSH to one switch, 47 00:04:05,230 --> 00:04:11,590 the one that has the management IP address and not have to telnet or SSH to multiple switches. 48 00:04:11,990 --> 00:04:19,160 There's one configuration file which is included across all the physical switches spanning tree, CDP, 49 00:04:19,180 --> 00:04:23,020 VTP or running on one switch, not on multiple switches. 50 00:04:23,530 --> 00:04:29,320 The ports on each physical switch appear to be part of the same logical switch. 51 00:04:29,330 --> 00:04:35,290 In other words, you may have four physical switches, each having their own physical ports. 52 00:04:35,620 --> 00:04:39,550 But logically you've got one switch with all of those ports. 53 00:04:40,030 --> 00:04:45,100 There's one Mac address table that references all the ports on all physical switches. 54 00:04:45,700 --> 00:04:52,330 There are some additional advantages, but the moral of the story is you have a single virtual switch 55 00:04:52,330 --> 00:04:57,130 that you managing rather than four separate, discrete switches. 56 00:04:57,850 --> 00:05:04,060 Now, to connect the switches together, you use special hardware ports called stacking ports. 57 00:05:04,570 --> 00:05:05,680 Again, there are different 58 00:05:05,680 --> 00:05:13,060 Cisco Technologies, we've got, as an example, Cisco FlexStack and FlexStackPlus stacking technology. 59 00:05:13,660 --> 00:05:19,720 You would, as an example, insert a stacking module in each switch and then connect them with a stacking 60 00:05:19,720 --> 00:05:20,200 cable. 61 00:05:20,830 --> 00:05:29,680 Different stacking technologies work on different switches, so FlexStack and FlexStackPlus all supported 62 00:05:29,680 --> 00:05:40,300 by 2960 switches such as the 2960S, 2960R, 2960X and 2960XR switch families. 63 00:05:41,450 --> 00:05:47,210 3750 switches, which are all the switches support Stackwise and what are the stacking 64 00:05:47,210 --> 00:05:50,900 cables do is they form a ring between the switches. 65 00:05:51,960 --> 00:05:57,240 In other words, the switches are connected in a series with a last switch connected back to the first 66 00:05:57,240 --> 00:06:04,260 switch as shown in this topology, by using full duplex on each link, the stacking modules and cables 67 00:06:04,260 --> 00:06:08,940 create two paths to forward data between the physical switches in the stack. 68 00:06:09,630 --> 00:06:15,360 The switches use these connections to communicate between the switches in the stack, as well as to 69 00:06:15,360 --> 00:06:18,960 forward frames and perform other overhead functions.