Hierarchical Network Design

17 minutes
Share the link to this page
You need to have access to the item to view this lesson.
One-time Fee
List Price:  $139.99
You save:  $40
List Price:  €129.04
You save:  €36.87
List Price:  £110.53
You save:  £31.58
List Price:  CA$191.55
You save:  CA$54.73
List Price:  A$211.24
You save:  A$60.35
List Price:  S$189
You save:  S$54
List Price:  HK$1,093.20
You save:  HK$312.36
CHF 92.16
List Price:  CHF 129.04
You save:  CHF 36.87
NOK kr1,065.65
List Price:  NOK kr1,491.95
You save:  NOK kr426.30
DKK kr688.26
List Price:  DKK kr963.60
You save:  DKK kr275.33
List Price:  NZ$228.66
You save:  NZ$65.33
List Price:  د.إ514.18
You save:  د.إ146.92
List Price:  ৳16,401.27
You save:  ৳4,686.41
List Price:  ₹11,627.39
You save:  ₹3,322.35
List Price:  RM658.58
You save:  RM188.18
List Price:  ₦204,567.22
You save:  ₦58,451.95
List Price:  ₨38,898.11
You save:  ₨11,114.54
List Price:  ฿5,122.47
You save:  ฿1,463.66
List Price:  ₺4,507.41
You save:  ₺1,287.92
List Price:  B$723.35
You save:  B$206.68
List Price:  R2,578.54
You save:  R736.77
List Price:  Лв252.74
You save:  Лв72.21
List Price:  ₩191,324.33
You save:  ₩54,668
List Price:  ₪513.44
You save:  ₪146.71
List Price:  ₱8,148.04
You save:  ₱2,328.18
List Price:  ¥21,970.72
You save:  ¥6,277.79
List Price:  MX$2,337.80
You save:  MX$667.99
List Price:  QR509.87
You save:  QR145.68
List Price:  P1,899.76
You save:  P542.82
List Price:  KSh18,247
You save:  KSh5,213.80
List Price:  E£6,597.01
You save:  E£1,884.99
List Price:  ብር8,034.09
You save:  ብር2,295.62
List Price:  Kz118,761.77
You save:  Kz33,934.36
List Price:  CLP$127,220.83
You save:  CLP$36,351.40
List Price:  CN¥995.39
You save:  CN¥284.42
List Price:  RD$8,234.17
You save:  RD$2,352.78
List Price:  DA18,814.36
You save:  DA5,375.91
List Price:  FJ$317.44
You save:  FJ$90.70
List Price:  Q1,086.15
You save:  Q310.35
List Price:  GY$29,254.27
You save:  GY$8,358.96
ISK kr13,791.40
List Price:  ISK kr19,308.52
You save:  ISK kr5,517.11
List Price:  DH1,394.74
You save:  DH398.52
List Price:  L2,480.57
You save:  L708.78
List Price:  ден7,947.19
You save:  ден2,270.78
List Price:  MOP$1,125.23
You save:  MOP$321.51
List Price:  N$2,569.81
You save:  N$734.28
List Price:  C$5,146.35
You save:  C$1,470.49
List Price:  रु18,594.99
You save:  रु5,313.23
List Price:  S/522.73
You save:  S/149.36
List Price:  K543.39
You save:  K155.26
List Price:  SAR525.04
You save:  SAR150.02
List Price:  ZK3,731.94
You save:  ZK1,066.34
List Price:  L642.02
You save:  L183.44
List Price:  Kč3,179.50
You save:  Kč908.49
List Price:  Ft49,522.43
You save:  Ft14,150.27
SEK kr1,086.34
List Price:  SEK kr1,520.92
You save:  SEK kr434.58
List Price:  ARS$124,480.29
You save:  ARS$35,568.34
List Price:  Bs966.19
You save:  Bs276.07
List Price:  COP$541,025.52
You save:  COP$154,589.76
List Price:  ₡71,687.89
You save:  ₡20,483.71
List Price:  L3,455.41
You save:  L987.33
List Price:  ₲1,051,715.43
You save:  ₲300,511.59
List Price:  $U5,389.29
You save:  $U1,539.90
List Price:  zł549.70
You save:  zł157.07
Already have an account? Log In


All right, welcome back. Now big word that nobody likes to pronounce. I'm going to try it hierarchical. I think I got it now or design. Okay, that is what we're going to talk about in this particular lecture. Okay.

Now, you'll see, you know, maybe do you know, campus networks are enterprise networks because they're huge, right? They got multiple buildings all over the place. Some of them, you know, you got some, I mean, look at all these universities, I believe universities are like city blocks, you know. So they're huge. All right. They consist of, of many, many, many, many, many lands, all right.

All in different buildings. There's usually one geographic located in the same area, of course, all right there. Some are in the same area, but they cover friggin 10 1520 blocks, right, not to mention the dorms and all these different things that they have in their restaurants, all the different things like a city and their company normally owns the network infrastructure, obviously Ethernet and wireless and I can tell you this, when I was at USM, they brought in wireless towards the end when I was leaving, ah, not a very easy task where you have a lot of things that interfere with wireless you we know this, but they brought wireless in so it's their own. No, no, that's not their own. And if you know this from the CCNA here's what Come on. We're the De Morgan's right.

Passive demark that's not yours. You don't own that only your CPE your customer premise equipment, right? But usually your lands are yours, okay. But I didn't want to cross when they no longer are yours. Okay, we know this. We know this.

But think about it. It is exciting. Think about it. You're going across buildings upon buildings upon buildings and within the same building. You have, you know, I have up to five or nine floors, okay? And you have to be able to network this and it's got to be transparent To the user to the student who's they're actually trying to find something further research or their paper for whatever the kids are going inside the library, they're cramming for their, their finals and all that stuff, you know, those now are being bogged down.

So you got to make sure now definitely in enterprise or campus networks, which both are the same, I you would need to make sure that you know, you're running no less than gigabit within and within even within the classrooms. All right, you can be running gigabit 90 100 megabits per second to the classrooms or anything like that. Everything's gonna be running gigabit. And anything above that. When you go into the distribution layers, or the core layers. Now we're talking about Metro ease 10 gigs, we get into that type of speed, you have to you have to buy more time by a hierarchy, a hierarchy.

All right, so how are we doing now? One way or the way that Cisco saying that you should look when you're designing your hierarchy is based on traffic flow. That's why I say it's a vital part of the campus network design. emphasis should be facilitating on the overall design related to known study or predictable traffic flows, or are these based on what a worst of all traffic going? Which way is it going? Where is the most traffic, this is where I need to focus.

And we're going to put my best switches, I'm going to put you know, the greater speeds, okay, I'm going to less amount of, let's say, protocols like access list or things of that nature or any kind of tunneling or any like that you need to kind of not burn so much the processing power, because if the traffic flow is going that way, you have made sure that those things are not going to burn the switch. And then on top of that, yeah, the data going through there. So traffic flow should be when you're creating this hierarchy. In the network, the way the path is going the way the data of the path flow is going, that should be your focal point to creating the network, okay? Now, let's go back to the beginning of time. hubs.

You're not gonna have any scalability, or any efficiency or any speed or nothing. If you have hubs in your network, this is ridiculous. This is a shared collision domain. We know this, I don't need to go into this. We know this. Okay.

The more you plug into that hub, the less bandwidth you're going to have. It does not segment logically physically or anything. It's going to be a huge mess. It was going to be low bandwidth, collisions beyond belief. Don't even look at this because you're not even going to do that. Okay, well, you are going to do this.

You are going to have switches does now this is a layer two switch where we get into private collision domains. Now Yeah, the full bandwidth Coming out of each port, which is great. Okay, but you still have that one broadcast. So we need to address that. We need to address that. Because if everybody's on VLAN, one below, we're creating one huge broadcast domain.

I don't care how good STP Spanning Tree Protocol is. All right, it will you'll run into an issue, you run into an issue. Well, here's my small network, okay, and a small number like this. No big deal. Okay, being on one. Everybody has in their own collision domains.

All right. We are one broadcast, but not a big deal. People are printing their computers, not a biggie. But you need to make sure if you're going to scale and that's one thing when you're a network engineer or network person, period, if you're going into a business, you need to say okay, is this business going to grow? You need to think about things like this, so you can get the right type of equipment to begin with. Okay, are we going to need firewalls, every I think every company should have some sort of firewall, either software or hardware.

Depending I guess, the size, not really the size of the company, but more of the information that the company is holding. It could be a very small attorneys firm. And that's very confidential information. They should not for a wall, small dentist office, a small doctor's office, small psychiatry office, they have confidential information, they should have some good security in there. You should include firewalls in there. Okay.

So, but as far as segmenting it up to, you know, how big is this psychiatry or dental or whatever office you're building that, you know, are they going to have really more than three or four or five or 1015 employees within their particular company. So you need to be aware. So normally, I always get, you know, depending on the company, but normally it's a 24 port switch. Okay, I'm going to go to 24 port switch, and now I look to see okay, I'm going to get to layer three. I'm going to get to layer two. So Am I going to build that hierarchy going from layer two to layer three, okay, and with the layer three handler routing, alright, and put everything on there.

So you got to see, but definitely, this is your be your second step, going into private collision domains because you are segmenting it now, physically, you're getting rid or you're increasing the number of collision domains while decreasing the amount of collision. So then you can increase the amount of bandwidth. Awful, but we've said this before the past with a CCNA. But now we get into this little particular situation right here. No biggie. We've seen this before.

I'll be I'll be like, well, I'm willing to Yeah, all right, what happens on this demon stays on that VLAN it's not gonna go across the router and come down here. It's not gonna happen. Now as VLAN one wants to talk to VLAN two, then that's what we have a layer three device. It could be a layer three switch, or it could be a router as you see here, and then you would configure the right protocol. In this case, this would be an it'll, it'll to on cue protocol on the switch on the router, this would be trunk ports. And then you'll go ahead and send information across to the VLAN.

But they'll go through the router and then back down. They will not go just to the switch. So now your segment one broadcast domain here, you know, one broadcast domain here. Okay. And then you still have Well, there's one broadcast to broadcast. This is VLAN.

One, this will be part of your one over here as well. So this will be the same broadcast as this. Okay, you're not going to put this in a different broadcast domain VLAN, one belongs to broadcast domain, and you have 12345 collision domains in here. But again, inter VLAN communication, so now you're actually logically segmenting everything using VLANs. To increase the bandwidth for everybody. You're actually taking away the collision domains and you're taking away bro caste domains.

Okay, you're, you're increasing, I'm sorry, you increasing the number of broadcast domains. All right, which is a good thing, because now you're decreasing the amount of broadcast in a particular segment and a particular subnet. But what are you doing? You're now going from this switch to the router, not a router has to process information and then bring it back down to the switch and then set it to where it needs to go. So let's say this server right here, we have a lot of information in this server, and we have to access that constantly. So if that's the case, then we're creating a lot of traffic flow.

So we better make sure that this backbone right here, that's why I say you can't have anything less than gigabit speed on these things. We have gigabit NICs, gigabit cable, gigabit switch, you know, the router has to have a gigabit interface, all that stuff so they can run nice and easy. Okay, and now we run into issues. What do we fail to see here? What do we fail to see here? We're done.

Done. See, we're creating a hierarchy. Okay, here's our access layer switches. Here's a distribution switch is gonna have a core switch. Not yet. All right, but we have a distribution switch.

And we have a core, we have an access switch. All right, and we have some redundancy, hey, one link goes down, we just go down the other way. And this guy, one link goes down, you're done. You're not going anywhere. Okay, here at least you have some redundancy going on here. And this little triangle, and as you can see through the lights, that SCP is doing its job saying, Hey, we're blocking that particular port because we don't want any loops to happen in our switching layer.

Okay, so we're getting better. We're getting better. Now we have a hierarchy, but we also have some little bit of redundancy. There's got to be careful what type of devices is that we're using that which way the traffic is going to go and is being constantly accessed, that we need to consider Because remember the magical number 80% 80% of all your resources should be local to your segment. Right? You're not going to print if you're on the first floor to printers on the ninth floor.

I just wouldn't make any sense. Okay, and don't be cheesy either. And have one every other floor. Wherever you're on two floors don't do that. We all need exercise, but now I work. Okay.

All right now there we go. All right. Now we got our core. Got our course which Yeah, course which is interested in what speed up? Right? We want to bog this guy who's going from one campus to the other.

One building to the other. His concern is speed. That's it, nothing else. Here now these will be distribution. Okay. And then these are access switches down here that go to our end devices.

Okay. So these are now become your workhorses. made your access list create your VLAN. Well, you can create VLANs down here, okay? And your protocols and protocols will go all over the place. But depending if you're doing tunneling, or how do you choose, because remember, we're going to have a collapse architecture as well.

All right. But here again, we don't have any redundancy. We have a great hierarchy. Why don't we go from our access switches? our computers are down here, we go to our access switch, where do you want to go? Or you want to go to another particular switch over here?

So you go to the distribution switch, the switches, take a look at what what is switches, look at their Mac tables, right then from their Mac tables, they go to their course words, of course, which makes the layer three more likely it's going to go ahead and well, these are all layer three switches that are going to go ahead and make a decision where you need to go boom, and it sends you out this way. Okay, to where you need to go. Again, traffic flow, which way is it going? So our Cisco through their model, and it's funny that you mentioned that now in the CCNP when they The CCNA is so focused on the collapse architecture. So let's go through our model. We'll talk about the access layer, low cost per switch port, high port density, scalable uplinks, high availability, ability to reverse networks, and security features and quality of service.

Okay? Not to say cheap or frugal, but these are these even though these are your workhorses really, they're doing a lot of work. You're not gonna have to have that much on there, but you will have high port density, okay, because all the information is going through there. So make sure that they have what they need in order for information to travel, the processing power, the RAM, the cabling, all that stuff. They need to have high availability, all right, they cannot be burned down, they cannot go down because that's that's leading to your end devices. Okay.

Ability to conversion network sources, that is data or voice. You can have phones connected, that could be a switch for phones. You Voice VLANs on that thing, alright, so one whole switch is dedicated to phones. So you gotta be high availability, more lawyers. And you got to make sure that it goes to invert whatever source is using. All right security features and quality of servers, you can tweak the quality of servers or right depending on what kind of devices you have.

And then security security goes without saying, it goes on. Really, to me, it goes on all three layers, you got to have security configured on your switches or on your routers, on your servers on everything that you have. You have to have redundant security layers throughout your entire network. Okay, distribution layer. This is where your aggregation of multiple layers which is having high layer three routing throughput in your packet handling, okay, you will have all the routing that's going on to be able to take care of that information, security and policy based connectivity functions where you create your policies. Again, high availability, their distribution, the Here's your end device going to the access switch.

And then you're not going to have your distribution switches available for that information to go on to the core, because you need to get to the other side. She was features, obviously. And then scalability and redundant, high speed links to the core, which we didn't have, which we didn't have. May I be high speed, but there's no redundancy, no redundancy, we have a hierarchy, like no redundancy. That's a problem. Okay, here we have the core layer, very high layer three routing throughput domains, that's your LAN.

We'll talk about why now may not be when maybe an interface going to another building within your land, but it's also going out to your land. So that's where your Metro II would be, let's say, all right, so definitely need a very, very high throughput in that particular switch. No costly, unnecessary packet manipulation, such as ACLs packet filtering, don't put stuff on that switch that doesn't need to be there. Okay. Because you're burning, don't do that. ACLs okay, but there's other types of security that you can use, okay, within that switch or putting a firewall in front of it or what have you, or behind or what, something that you can go ahead and say, Okay, we'll put no no policies, but we'll definitely go ahead and use some sort of security because I do believe in putting security on there and it says, like ACLs, okay, don't burn it with 200 or 2000 ACLs.

But put something on there. redundancy and resilience for high availability, which we didn't have, which we did not have. Okay, so you need to think about redundancy, redundancy. QoS. Once again, I I'm going with this Oh, here we go. When the distribution and core layers are combined into a single layer switch, it becomes a collapsed core.

Okay, again, distribution and core layers are combined into a single layer of switches. It becomes a collapsed core. Okay, everybody what that means is your distribution and your access as one block east. So you have a block here, block here, number blocks segment here segment here, well distribution and access layer switches, and then they're all connected to your core. Okay, so they're working together. So here's what the system will do their model that they're talking about comes in with all the different layers, right?

Core distribution and access. But again, you have a great hierarchy, but you have no none whatsoever. None whatsoever. redundancy, one link goes down, you're done. Right? And that's a problem.

You don't need that. We don't want that and has no engineers. We need to make sure that we have that same connection.

Sign Up


Share with friends, get 20% off
Invite your friends to LearnDesk learning marketplace. For each purchase they make, you get 20% off (upto $10) on your next purchase.