1 1 00:00:00,870 --> 00:00:04,180 In this lesson, we're going to talk about network topologies. 2 2 00:00:04,180 --> 00:00:06,310 And when we talk about typologies for networks, 3 3 00:00:06,310 --> 00:00:08,800 we're going to be talking about it in one of two ways. 4 4 00:00:08,800 --> 00:00:11,240 First, we can talk about it physically. 5 5 00:00:11,240 --> 00:00:13,460 How are these devices physically cabled 6 6 00:00:13,460 --> 00:00:16,430 and connected together using various types of media. 7 7 00:00:16,430 --> 00:00:19,350 And the second is how it is done logically. 8 8 00:00:19,350 --> 00:00:21,870 We talk about logically, we're talking about how the traffic 9 9 00:00:21,870 --> 00:00:24,310 is actually going to flow in that network. 10 10 00:00:24,310 --> 00:00:26,100 So, here on the screen, you can see 11 11 00:00:26,100 --> 00:00:29,040 I have a logical diagram of what a network looks like. 12 12 00:00:29,040 --> 00:00:31,520 You could see where the workstations are, the routers, 13 13 00:00:31,520 --> 00:00:34,260 the switches, all of that stuff, but this is not 14 14 00:00:34,260 --> 00:00:37,070 how the network actually looks in the real world. 15 15 00:00:37,070 --> 00:00:40,050 For example, that Windows 7 machine in the upper left 16 16 00:00:40,050 --> 00:00:41,810 might be on the third floor to the building, 17 17 00:00:41,810 --> 00:00:44,230 while the Windows 2012 machine on the upper right 18 18 00:00:44,230 --> 00:00:46,440 may be all the way down in our basement. 19 19 00:00:46,440 --> 00:00:48,420 You really don't know based on this diagram 20 20 00:00:48,420 --> 00:00:50,280 where these things are physically, 21 21 00:00:50,280 --> 00:00:52,560 because this is a logical topology, 22 22 00:00:52,560 --> 00:00:55,060 I'm just concerned with the way the data flows, 23 23 00:00:55,060 --> 00:00:57,290 not the way it's actually cabled. 24 24 00:00:57,290 --> 00:00:59,730 Now, here we're going to learn how to read these diagrams 25 25 00:00:59,730 --> 00:01:01,430 throughout the rest of this course and what each of these 26 26 00:01:01,430 --> 00:01:04,350 icons mean, but right now that's not important. 27 27 00:01:04,350 --> 00:01:06,040 I just want you to realize that there's a difference 28 28 00:01:06,040 --> 00:01:08,530 between a logical and a physical topology 29 29 00:01:08,530 --> 00:01:11,270 and that difference is one is focused on the logical 30 30 00:01:11,270 --> 00:01:13,560 or network flow and the other is focused 31 31 00:01:13,560 --> 00:01:16,310 on the physical layout of the cabling. 32 32 00:01:16,310 --> 00:01:18,610 So, what are some of these options that we have 33 33 00:01:18,610 --> 00:01:20,210 for different topologies? 34 34 00:01:20,210 --> 00:01:22,310 Well, the first one we have is what's known 35 35 00:01:22,310 --> 00:01:23,940 as a bus topology. 36 36 00:01:23,940 --> 00:01:26,530 This is where you use a single cable that runs the length 37 37 00:01:26,530 --> 00:01:29,490 of the entire area that needs network connectivity. 38 38 00:01:29,490 --> 00:01:31,940 Each machine, whether 't's a laptop or a desktop 39 39 00:01:31,940 --> 00:01:34,350 or a server, will then tap into that cable 40 40 00:01:34,350 --> 00:01:37,210 using either a T-connector or a vampire tap. 41 41 00:01:37,210 --> 00:01:39,680 Now, you may be wondering what's a vampire tap, 42 42 00:01:39,680 --> 00:01:42,060 well, it was an old way of connecting networks. 43 43 00:01:42,060 --> 00:01:44,760 Essentially, we would take a big thick metal cable 44 44 00:01:44,760 --> 00:01:46,850 and run it down the room where all the computers 45 45 00:01:46,850 --> 00:01:49,100 were to be connected and then each computer 46 46 00:01:49,100 --> 00:01:52,530 would have a clamp that would bite into the cable 47 47 00:01:52,530 --> 00:01:54,330 and it would actually make the connection to the network 48 48 00:01:54,330 --> 00:01:56,800 that way and that's why we called it a vampire tap. 49 49 00:01:56,800 --> 00:01:58,890 Now, this is a very old technology, it was used 50 50 00:01:58,890 --> 00:02:02,070 back in the eighties and nineties, we don't use this today, 51 51 00:02:02,070 --> 00:02:05,240 but, we still do have some bus typologies out there. 52 52 00:02:05,240 --> 00:02:07,970 We don't use them very often, but they do still exist. 53 53 00:02:07,970 --> 00:02:11,130 If you see a bus topology, all the devices on this cable 54 54 00:02:11,130 --> 00:02:14,030 would form what's known as a single collision domain. 55 55 00:02:14,030 --> 00:02:16,490 This means, as you can see here, there are six 56 56 00:02:16,490 --> 00:02:18,300 different devices here trying to talk. 57 57 00:02:18,300 --> 00:02:20,290 If they all try to talk at the same time, 58 58 00:02:20,290 --> 00:02:22,230 you just have a collision, because they're all sharing 59 59 00:02:22,230 --> 00:02:25,480 the same cable, so they'd have to instead take turns. 60 60 00:02:25,480 --> 00:02:27,220 The next topology we're going to discuss 61 61 00:02:27,220 --> 00:02:29,870 is known as a ring topology and this uses 62 62 00:02:29,870 --> 00:02:32,690 a single cable like a bus does, but instead of it being 63 63 00:02:32,690 --> 00:02:36,010 in a straight line, it's going to run in a complete circle. 64 64 00:02:36,010 --> 00:02:38,580 Each device in the ring can then talk on that cable, 65 65 00:02:38,580 --> 00:02:40,350 but again, they have to wait their turn 66 66 00:02:40,350 --> 00:02:42,080 or you're going to have a collision. 67 67 00:02:42,080 --> 00:02:44,150 Data would travel around this in a single way, 68 68 00:02:44,150 --> 00:02:46,090 either clockwise or counterclockwise, 69 69 00:02:46,090 --> 00:02:48,610 depending on the configuration of your network. 70 70 00:02:48,610 --> 00:02:50,730 Now, we had these, because of that collision 71 71 00:02:50,730 --> 00:02:53,610 we wanted to overcome, what we did was we had what's called 72 72 00:02:53,610 --> 00:02:55,720 a token ring and a token was just 73 73 00:02:55,720 --> 00:02:57,620 an electronic tag essentially, 74 74 00:02:57,620 --> 00:03:00,870 that was passed around logically from computer to computer, 75 75 00:03:00,870 --> 00:03:02,400 as they were going to talk. 76 76 00:03:02,400 --> 00:03:04,180 This would allow them to take turns. 77 77 00:03:04,180 --> 00:03:06,140 So, if you think back to when you're in elementary school 78 78 00:03:06,140 --> 00:03:08,740 and everyone's sitting around a circle and there's 20 kids 79 79 00:03:08,740 --> 00:03:11,720 and one teacher, the teacher may have had the talking stick 80 80 00:03:11,720 --> 00:03:13,200 and they gave the stick to a child 81 81 00:03:13,200 --> 00:03:15,110 and that child can then talk, then they would take 82 82 00:03:15,110 --> 00:03:16,580 that stick away and give it to a different child 83 83 00:03:16,580 --> 00:03:18,020 and then that child could talk. 84 84 00:03:18,020 --> 00:03:20,060 And this way, everyone could hear what everybody else 85 85 00:03:20,060 --> 00:03:22,590 was saying, without any collisions or anyone talking 86 86 00:03:22,590 --> 00:03:26,090 over each other, that's the idea of a token ring network. 87 87 00:03:26,090 --> 00:03:27,790 This eliminates those collisions that you had 88 88 00:03:27,790 --> 00:03:29,760 back on the bus network. 89 89 00:03:29,760 --> 00:03:32,610 Now, when you had these rings, they still were vulnerable 90 90 00:03:32,610 --> 00:03:34,090 to being broken. 91 91 00:03:34,090 --> 00:03:36,700 If you cut the cable, that would take down the network, 92 92 00:03:36,700 --> 00:03:38,800 so there was no redundancy. 93 93 00:03:38,800 --> 00:03:41,180 Now, there is something known as a FDDI ring, 94 94 00:03:41,180 --> 00:03:43,370 which is a fiber distribution network. 95 95 00:03:43,370 --> 00:03:47,050 A FDDI ring actually uses two rings one on top of the other, 96 96 00:03:47,050 --> 00:03:48,930 one operates in a clockwise direction 97 97 00:03:48,930 --> 00:03:51,860 and the other operates in a counterclockwise direction. 98 98 00:03:51,860 --> 00:03:54,530 When you use a FDDI ring, you do get redundancy, 99 99 00:03:54,530 --> 00:03:56,400 because if one of those rings is broken, 100 100 00:03:56,400 --> 00:03:59,010 the other one would take over the network load. 101 101 00:03:59,010 --> 00:04:03,160 Now, for the network plus exam, if you see the word ring, 102 102 00:04:03,160 --> 00:04:05,410 I want you to think about a FDDI ring, 103 103 00:04:05,410 --> 00:04:07,440 those fiber optic distribution rings, 104 104 00:04:07,440 --> 00:04:10,440 because these days, that's the only type of ring topology 105 105 00:04:10,440 --> 00:04:12,390 we really use in our networks. 106 106 00:04:12,390 --> 00:04:14,780 For the exam, anytime they ask you a question 107 107 00:04:14,780 --> 00:04:16,880 and ring is an option, I want you to think 108 108 00:04:16,880 --> 00:04:18,910 ring equals redundancy. 109 109 00:04:18,910 --> 00:04:21,660 Now, in the real world, if you had a standard token ring 110 110 00:04:21,660 --> 00:04:24,280 with a single ring, there's no redundancy there, 111 111 00:04:24,280 --> 00:04:26,590 but for the exam, I'm telling you, 112 112 00:04:26,590 --> 00:04:28,460 they're talking about FDDI rings. 113 113 00:04:28,460 --> 00:04:31,570 So, when you see ring, think about redundancy. 114 114 00:04:31,570 --> 00:04:33,420 You might get a question that says, 115 115 00:04:33,420 --> 00:04:35,650 which of these typologies is redundant 116 116 00:04:35,650 --> 00:04:38,010 and they're going to give you options like star and bus 117 117 00:04:38,010 --> 00:04:41,270 and ring and the answer is going to be ring. 118 118 00:04:41,270 --> 00:04:43,720 All right, I think I stress that and repeat that enough, 119 119 00:04:43,720 --> 00:04:45,740 you're going to remember, ring equals redundancy 120 120 00:04:45,740 --> 00:04:47,330 for the exam now, right? 121 121 00:04:47,330 --> 00:04:49,370 All right, let's move on to the next topology. 122 122 00:04:49,370 --> 00:04:52,270 The next one we have is called a star topology. 123 123 00:04:52,270 --> 00:04:54,340 You'll notice here, all of the outlying machines 124 124 00:04:54,340 --> 00:04:56,420 are talking to this central point. 125 125 00:04:56,420 --> 00:04:58,020 This central point is normally going to be something 126 126 00:04:58,020 --> 00:04:59,230 like a switch. 127 127 00:04:59,230 --> 00:05:01,170 Now, this switch is where all of the things 128 128 00:05:01,170 --> 00:05:03,230 are going to connect back to and this happens 129 129 00:05:03,230 --> 00:05:05,240 in most of our networks these days. 130 130 00:05:05,240 --> 00:05:07,940 You can use it with fiber or copper or even wireless, 131 131 00:05:07,940 --> 00:05:11,450 if you wanted to and this is going to be using a star topology. 132 132 00:05:11,450 --> 00:05:13,580 We call it a star because everything is bursting out 133 133 00:05:13,580 --> 00:05:15,090 like a star pattern. 134 134 00:05:15,090 --> 00:05:16,770 But, the problem with this is that you have 135 135 00:05:16,770 --> 00:05:19,610 this one central device and so if I have a switch here 136 136 00:05:19,610 --> 00:05:22,650 in the middle and the switch fails because it loses power, 137 137 00:05:22,650 --> 00:05:25,010 or we cut the cable or something like that, 138 138 00:05:25,010 --> 00:05:26,990 the entire network is going to fail, 139 139 00:05:26,990 --> 00:05:29,010 because it's a single point of failure. 140 140 00:05:29,010 --> 00:05:31,040 So, remember with a star that you always have 141 141 00:05:31,040 --> 00:05:33,580 this single point of failure, even though these are 142 142 00:05:33,580 --> 00:05:36,890 very, very common to use and very inexpensive to use, 143 143 00:05:36,890 --> 00:05:38,640 we do have this single point of failure 144 144 00:05:38,640 --> 00:05:40,700 sitting right there in the middle. 145 145 00:05:40,700 --> 00:05:43,400 Now, the next topology, we have tries to overcome that 146 146 00:05:43,400 --> 00:05:45,480 and it's called a hub-and-spoke. 147 147 00:05:45,480 --> 00:05:47,290 Now, a hub-and-spoke to topology is used 148 148 00:05:47,290 --> 00:05:49,500 for connecting multiple sites together. 149 149 00:05:49,500 --> 00:05:51,270 The reason we call it a hub-and-spoke, 150 150 00:05:51,270 --> 00:05:53,380 is if you think about it, the way the airlines operate, 151 151 00:05:53,380 --> 00:05:55,320 it looks like a hub-and-spoke. 152 152 00:05:55,320 --> 00:05:57,560 These usually have hubs, like for instance where I live 153 153 00:05:57,560 --> 00:06:00,017 in Baltimore, there's a hub for Southwest Airlines 154 154 00:06:00,017 --> 00:06:03,280 and if I wanted to go to Connecticut and go to California, 155 155 00:06:03,280 --> 00:06:05,410 you'd actually fly from Connecticut to Baltimore, 156 156 00:06:05,410 --> 00:06:08,430 which is the hub and then go out over to California. 157 157 00:06:08,430 --> 00:06:10,760 This allows them to centralize their operations 158 158 00:06:10,760 --> 00:06:11,950 into hub cities. 159 159 00:06:11,950 --> 00:06:14,750 There's usually each airline has two or three or four hubs 160 160 00:06:14,750 --> 00:06:16,710 and all the flights go through those. 161 161 00:06:16,710 --> 00:06:19,000 So, because you go to the hub first and then you go out 162 162 00:06:19,000 --> 00:06:20,280 to the spoke cities, 163 163 00:06:20,280 --> 00:06:22,690 this is the same concept with the network. 164 164 00:06:22,690 --> 00:06:25,120 It's very similar to the star, but there's multiple 165 165 00:06:25,120 --> 00:06:28,000 WAN links and multiple places that these hubs can exist. 166 166 00:06:28,000 --> 00:06:30,100 So, it's almost like a hybrid of taking those stars 167 167 00:06:30,100 --> 00:06:31,630 and connecting them together. 168 168 00:06:31,630 --> 00:06:33,020 So, as you can see here on the screen, 169 169 00:06:33,020 --> 00:06:35,810 I have Denver and Los Angeles as the hubs in my network 170 170 00:06:35,810 --> 00:06:37,800 and everyone else is going to be a spoke. 171 171 00:06:37,800 --> 00:06:39,910 So, if I wanted to go from Atlanta to Seattle, 172 172 00:06:39,910 --> 00:06:43,280 I have to go from Atlanta to Denver and Denver to Seattle. 173 173 00:06:43,280 --> 00:06:45,730 Now, if I wanted to go from Minneapolis to San Francisco, 174 174 00:06:45,730 --> 00:06:48,660 I could go from Minneapolis to Denver and Denver to LA 175 175 00:06:48,660 --> 00:06:50,400 and LA to San Francisco. 176 176 00:06:50,400 --> 00:06:53,250 No matter which way I go, I have to go through either LA 177 177 00:06:53,250 --> 00:06:55,840 or Denver, because those are the hub nodes. 178 178 00:06:55,840 --> 00:06:58,160 It's not redundant fully, because if one of those 179 179 00:06:58,160 --> 00:07:00,610 central offices like Denver or LA fails, 180 180 00:07:00,610 --> 00:07:02,920 we're going to lose large portions of our network, 181 181 00:07:02,920 --> 00:07:05,570 but it is better than a single star. 182 182 00:07:05,570 --> 00:07:08,040 Now, because of this, losing half of our network 183 183 00:07:08,040 --> 00:07:10,150 is a bad thing, right, this is a problem. 184 184 00:07:10,150 --> 00:07:13,140 If we're looking for full redundancy, we would want to go 185 185 00:07:13,140 --> 00:07:15,630 to what's called a full-mesh topology. 186 186 00:07:15,630 --> 00:07:19,270 Now, full-mesh is awesome when you talk about redundancy. 187 187 00:07:19,270 --> 00:07:21,290 Every single node or device is connected 188 188 00:07:21,290 --> 00:07:22,960 to every other node or device. 189 189 00:07:22,960 --> 00:07:26,320 This works really, really well if you have a small network, 190 190 00:07:26,320 --> 00:07:29,330 two or three machines, but as I start increasing 191 191 00:07:29,330 --> 00:07:31,340 the number of machines, it gets us get 192 192 00:07:31,340 --> 00:07:33,800 pretty crazy pretty quickly. 193 193 00:07:33,800 --> 00:07:36,100 Optimal routing is always available when you're dealing 194 194 00:07:36,100 --> 00:07:38,840 with full-mesh, because every machine can go direct 195 195 00:07:38,840 --> 00:07:41,930 with one jump over to the machine they want to get to. 196 196 00:07:41,930 --> 00:07:43,540 So, in this example you see, 197 197 00:07:43,540 --> 00:07:45,420 I have six machines on the network. 198 198 00:07:45,420 --> 00:07:47,990 Now, every single one of them is tied to each other, 199 199 00:07:47,990 --> 00:07:50,730 so, if I was doing this physically, I would need to have 200 200 00:07:50,730 --> 00:07:53,450 five network cards for each of these machines 201 201 00:07:53,450 --> 00:07:56,370 and five cables going from each of those machines. 202 202 00:07:56,370 --> 00:07:58,300 That's a lot of technology that I have to add 203 203 00:07:58,300 --> 00:08:00,970 to be able to connect these machines together, right? 204 204 00:08:00,970 --> 00:08:03,380 And so, if I counted up the number of black lines there, 205 205 00:08:03,380 --> 00:08:05,250 how many different connections are there? 206 206 00:08:05,250 --> 00:08:07,880 How many lines would I actually have to run? 207 207 00:08:07,880 --> 00:08:11,380 Well, there's going to be six machines times five cables each, 208 208 00:08:11,380 --> 00:08:14,790 which is 30, divided by two, which is 15. 209 209 00:08:14,790 --> 00:08:16,960 So, there's actually going to be, for six machines, 210 210 00:08:16,960 --> 00:08:19,420 it's going to take me 15 cables. 211 211 00:08:19,420 --> 00:08:21,940 Now, if I go to seven machines or 10 machines, 212 212 00:08:21,940 --> 00:08:24,280 it gets pretty crazy and gets really expensive 213 213 00:08:24,280 --> 00:08:26,050 really quickly and this is why 214 214 00:08:26,050 --> 00:08:28,300 you're never really going to see a full-mesh 215 215 00:08:28,300 --> 00:08:29,780 in a physical network. 216 216 00:08:29,780 --> 00:08:31,840 Instead, you're usually going to see full-mesh 217 217 00:08:31,840 --> 00:08:33,810 more as a logical method. 218 218 00:08:33,810 --> 00:08:36,310 If you have something like nuclear command and control, 219 219 00:08:36,310 --> 00:08:38,350 they may want to have full-mesh and they'll go through 220 220 00:08:38,350 --> 00:08:40,384 the expense, because they have a zero defect 221 221 00:08:40,384 --> 00:08:43,540 mentality there, but in your office and business networks, 222 222 00:08:43,540 --> 00:08:45,040 you're not going to be seeing full-mesh, 223 223 00:08:45,040 --> 00:08:48,010 it's just too expensive and too complicated. 224 224 00:08:48,010 --> 00:08:50,460 Now, what might you see in your business networks? 225 225 00:08:50,460 --> 00:08:53,270 Well, you might see a partial mesh network. 226 226 00:08:53,270 --> 00:08:56,330 Now, a partial mesh network is hybrid of the full-mesh 227 227 00:08:56,330 --> 00:08:57,710 and a hub-and-spoke. 228 228 00:08:57,710 --> 00:08:59,960 Using this design, we can provide optimal routing 229 229 00:08:59,960 --> 00:09:02,720 between some sites, but not all of the sites 230 230 00:09:02,720 --> 00:09:05,510 and so to get this right, you have to do a good survey 231 231 00:09:05,510 --> 00:09:07,590 and figure out where the busiest sites are 232 232 00:09:07,590 --> 00:09:09,330 and where the slow sites are. 233 233 00:09:09,330 --> 00:09:10,750 So, as you could see here on the diagram, 234 234 00:09:10,750 --> 00:09:12,470 if you start drawing your finger around, 235 235 00:09:12,470 --> 00:09:15,730 you can get to everywhere else, at least going through 236 236 00:09:15,730 --> 00:09:17,420 one or two different ways. 237 237 00:09:17,420 --> 00:09:18,710 Sometimes you're going to have to go through 238 238 00:09:18,710 --> 00:09:21,090 one of those central sites and if one of those central sites 239 239 00:09:21,090 --> 00:09:23,210 goes down, you can still get to everywhere else, 240 240 00:09:23,210 --> 00:09:25,970 because we have the right partial mesh set up. 241 241 00:09:25,970 --> 00:09:27,960 So, this partial mesh works really well 242 242 00:09:27,960 --> 00:09:30,340 and gives you that additional redundancy you're seeking. 243 243 00:09:30,340 --> 00:09:32,570 Now, to get it right though, you do have to understand 244 244 00:09:32,570 --> 00:09:35,180 your traffic patterns and know where to put those main hubs 245 245 00:09:35,180 --> 00:09:38,290 and it basically then becomes a modified hub-and-spoke, 246 246 00:09:38,290 --> 00:09:40,490 where we have more hubs than we did before 247 247 00:09:40,490 --> 00:09:42,810 and this allows us to have a better redundancy 248 248 00:09:42,810 --> 00:09:44,910 than a single hub-and-spoke configuration.