1 00:00:00,490 --> 00:00:02,770 In a similar way to our previous example. 2 00:00:02,770 --> 00:00:08,350 Let's assume that C replies to A so C sends a frame to the bridge. 3 00:00:08,440 --> 00:00:14,440 The bridge will read the source Mac address in the frame and then update its Mac address table with 4 00:00:14,440 --> 00:00:15,460 that information. 5 00:00:15,460 --> 00:00:18,910 So the bridge now knows that C is on Port three. 6 00:00:18,940 --> 00:00:26,980 As well as knowing that A is on port one because it learnt that from the previous frame and now unlike 7 00:00:27,010 --> 00:00:31,180 a hub, the bridge does not forward the frame out of all ports. 8 00:00:31,600 --> 00:00:38,230 The destination address in the frame is a the bridge knows that Mac address A is on port one so it only 9 00:00:38,230 --> 00:00:39,930 forwards the frame out of port one. 10 00:00:39,940 --> 00:00:43,630 The frame from C therefore only goes out of port one. 11 00:00:43,630 --> 00:00:49,810 It's not sent out of port two or port four because the bridge knows that A is on port one. 12 00:00:50,140 --> 00:00:51,380 So what does this mean? 13 00:00:51,400 --> 00:00:56,890 All subsequent frames from A and C will only use port one and three. 14 00:00:56,920 --> 00:01:02,350 In other words, if A sends another frame to see, it'll only go out a port three. 15 00:01:02,740 --> 00:01:08,440 This is because the MAC addresses of A and C are in the MAC address table and the bridge will forward 16 00:01:08,440 --> 00:01:11,530 traffic based on entries in the MAC address table. 17 00:01:11,560 --> 00:01:17,440 B and D are no longer receiving frames between A and C frames from C to A. 18 00:01:17,470 --> 00:01:22,090 Arriving on port three will go out port one and frames from A to C. 19 00:01:22,120 --> 00:01:25,060 Arriving on port one will be sent out of port three. 20 00:01:25,090 --> 00:01:33,160 Therefore, A and C can have a conversation independently of bad bad D on no longer receiving frames 21 00:01:33,160 --> 00:01:34,750 sent between A and C. 22 00:01:35,320 --> 00:01:39,970 The frames between A and C are contained between ports one and three. 23 00:01:40,090 --> 00:01:44,230 No bandwidth is used on port two and four. 24 00:01:44,230 --> 00:01:51,730 When traffic is sent between A and C, devices, B and D do not receive any frames sent between A and 25 00:01:51,730 --> 00:01:57,820 C and therefore avoid unnecessary processing of frames not destined to themselves. 26 00:01:58,030 --> 00:02:04,870 Bandwidth is being conserved, devices are not unnecessarily processing, traffic not destined to them, 27 00:02:04,870 --> 00:02:08,080 and thus bridges have major advantages over hubs. 28 00:02:08,910 --> 00:02:12,840 Over time, the bridge will learn where all Mac addresses are. 29 00:02:12,870 --> 00:02:19,050 So the bridge will learn that A is in port one, B's on port two, sees on port three and DD is on port 30 00:02:19,050 --> 00:02:19,620 four. 31 00:02:19,650 --> 00:02:27,110 That means that over time, B and D can have a conversation independently of A and C. 32 00:02:27,120 --> 00:02:29,940 The two conversations do not affect each other. 33 00:02:29,970 --> 00:02:35,490 Frames from each conversation do not interfere with the other conversation. 34 00:02:35,640 --> 00:02:40,560 Therefore, B and C can communicate at the same time as A and C. 35 00:02:41,960 --> 00:02:48,560 Now continuing with the advantages of bridges, each port is a different collision domain, so a collision 36 00:02:48,560 --> 00:02:50,840 on port one will not affect Port three. 37 00:02:51,050 --> 00:02:54,950 Each interface on a bridge is a separate collision domain. 38 00:02:55,130 --> 00:03:00,770 So in this example, we have one, two, three, four collision domains. 39 00:03:01,370 --> 00:03:07,700 If A and B were having a conversation and a collision took place on Port three, it will not affect 40 00:03:07,700 --> 00:03:12,920 A and B, they wouldn't even realize that there was a collision in the network. 41 00:03:13,310 --> 00:03:17,870 Now, in this topology, we have a hub connected to port four of the bridge. 42 00:03:17,900 --> 00:03:24,620 A hub is a single collision domain, so any collisions that take place on the hub will affect devices 43 00:03:24,620 --> 00:03:30,290 connected to the hub, but will not affect other devices elsewhere in the topology. 44 00:03:30,650 --> 00:03:37,550 So if there was a collision on the hub, it would affect host E and host DX, but it would not affect 45 00:03:37,550 --> 00:03:40,010 host A, C and B. 46 00:03:40,280 --> 00:03:45,800 The problem with collisions is that if a collision takes place, the devices have to back off for a 47 00:03:45,800 --> 00:03:49,820 random period of time and then they need to try and access the network again. 48 00:03:50,210 --> 00:03:56,510 So if these devices D and E are in a single collision domain, the bandwidth and throughput that they 49 00:03:56,510 --> 00:03:57,800 have is lower. 50 00:03:57,830 --> 00:04:06,110 Then these devices, which are in a separate collision domain by themselves A, C, and B have a dedicated 51 00:04:06,110 --> 00:04:06,890 link. 52 00:04:07,160 --> 00:04:11,420 They are on a single broadcast domain and single collision domain. 53 00:04:11,690 --> 00:04:19,940 D and E however, are sharing bandwidth because they connected to a hub host A, C and B are on separate 54 00:04:19,940 --> 00:04:21,260 collision domains. 55 00:04:21,680 --> 00:04:26,300 Now it's important to remember that a bridge is still a single broadcast domain. 56 00:04:26,300 --> 00:04:31,010 So if a sent a broadcast, it would be received by everyone in this topology. 57 00:04:31,250 --> 00:04:37,490 All devices will receive the broadcast, and in some cases that's a good thing, but in most cases it's 58 00:04:37,490 --> 00:04:38,120 not. 59 00:04:38,630 --> 00:04:43,670 In networking, we typically want to restrict or contain broadcast traffic. 60 00:04:43,670 --> 00:04:49,190 When there are too many broadcasts in the network, it can slow down all devices on the network and 61 00:04:49,190 --> 00:04:52,830 in the worst cases it will bring your network to its knees. 62 00:04:52,850 --> 00:04:58,010 In other words, your network will just break and not function if you have what's called a broadcast 63 00:04:58,010 --> 00:04:58,700 storm. 64 00:04:59,860 --> 00:05:05,980 Bridges once again process information in software rather than in hardware, and therefore tend to be 65 00:05:05,980 --> 00:05:12,310 slow in comparison to devices such as switches, which process frames in hardware. 66 00:05:12,520 --> 00:05:17,200 The number of ports on a bridge is also limited when compared to switches. 67 00:05:17,500 --> 00:05:23,140 In today's environments, switches have essentially replaced bridges, but it's good for you to realize 68 00:05:23,140 --> 00:05:26,410 that a bridge and a switch operate in a very similar way. 69 00:05:26,980 --> 00:05:31,150 So in summary, a bridge is a layer two device in the OSA model. 70 00:05:31,180 --> 00:05:33,850 In other words, it operates at the data link layer. 71 00:05:33,880 --> 00:05:39,670 It's more intelligent than a hub because it has a mac address table and it learns where Mac addresses 72 00:05:39,670 --> 00:05:46,090 are and then adds those MAC addresses to the MAC address table and can then make intelligent decisions 73 00:05:46,090 --> 00:05:51,820 on where to forward traffic based on the information learned and contained in the Mac address table. 74 00:05:51,880 --> 00:05:58,450 A hub is a physical device that simply repeats signals out of all ports, except the ports in which 75 00:05:58,450 --> 00:05:59,770 the traffic was received. 76 00:06:00,130 --> 00:06:05,500 A bridge will flood a frame out of all ports when it doesn't know where to send the frame. 77 00:06:05,770 --> 00:06:09,550 In other words, it hasn't learned where the destination Mac address is. 78 00:06:09,580 --> 00:06:13,120 It will also flood broadcasts out of all ports. 79 00:06:13,150 --> 00:06:17,460 So each port on a bridge is a separate collision domain. 80 00:06:17,470 --> 00:06:21,700 But a bridge is still a single broadcast domain. 81 00:06:22,730 --> 00:06:28,730 Switches are very similar to bridges as they both reside at layer two or the data link layer of the 82 00:06:28,910 --> 00:06:29,780 same model. 83 00:06:30,050 --> 00:06:35,510 The big advantage of switching when compared to bridging is that processing can be done in hardware 84 00:06:35,540 --> 00:06:40,460 using what are called ASICs or application specific integrated circuits. 85 00:06:40,730 --> 00:06:44,420 The number of ports supported by switches is also a lot higher. 86 00:06:44,450 --> 00:06:50,240 Hundreds of ports are supported on certain switches, whereas with bridges you were limited to a few 87 00:06:50,240 --> 00:06:50,970 ports. 88 00:06:50,990 --> 00:06:56,840 Switches are able to do this because processing is done in hardware and in actual fact, these days 89 00:06:56,840 --> 00:07:02,660 processing is done at wire speed, which means that there is no degradation in performance between two 90 00:07:02,660 --> 00:07:05,720 devices when they are connected via a switch. 91 00:07:05,750 --> 00:07:11,990 In other words, switches can move traffic from one port to another port at the same speed as if they 92 00:07:11,990 --> 00:07:13,280 weren't there. 93 00:07:13,610 --> 00:07:19,940 They can process and switch frames from one port to another port without slowing the frame down. 94 00:07:20,210 --> 00:07:26,810 So here's a quick comparison between switches and bridges switches process in hardware using a six bridges 95 00:07:26,810 --> 00:07:29,660 process in software and are therefore a lot slower. 96 00:07:29,690 --> 00:07:34,580 Switches support many ports bridges are limited in the number of ports that they support. 97 00:07:35,060 --> 00:07:38,720 Bridges have been replaced by switches in today's networks.