Modular Network Design

20 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:  €129.13
You save:  €36.89
List Price:  £110.03
You save:  £31.44
List Price:  CA$192.03
You save:  CA$54.87
List Price:  A$211.59
You save:  A$60.45
List Price:  S$189.02
You save:  S$54.01
List Price:  HK$1,093.71
You save:  HK$312.51
CHF 91.46
List Price:  CHF 128.06
You save:  CHF 36.59
NOK kr1,061.82
List Price:  NOK kr1,486.60
You save:  NOK kr424.77
DKK kr688.17
List Price:  DKK kr963.46
You save:  DKK kr275.29
List Price:  NZ$229.08
You save:  NZ$65.45
List Price:  د.إ514.18
You save:  د.إ146.92
List Price:  ৳16,474.51
You save:  ৳4,707.34
List Price:  ₹11,636.17
You save:  ₹3,324.85
List Price:  RM659.56
You save:  RM188.46
List Price:  ₦202,985.50
You save:  ₦58,000
List Price:  ₨39,085.72
You save:  ₨11,168.14
List Price:  ฿5,135.45
You save:  ฿1,467.37
List Price:  ₺4,513.94
You save:  ₺1,289.79
List Price:  B$720.20
You save:  B$205.78
List Price:  R2,573.67
You save:  R735.38
List Price:  Лв252.56
You save:  Лв72.16
List Price:  ₩191,210.73
You save:  ₩54,635.53
List Price:  ₪513.93
You save:  ₪146.84
List Price:  ₱8,148.95
You save:  ₱2,328.44
List Price:  ¥21,975.49
You save:  ¥6,279.16
List Price:  MX$2,337.44
You save:  MX$667.89
List Price:  QR512.15
You save:  QR146.33
List Price:  P1,899.46
You save:  P542.74
List Price:  KSh18,548.67
You save:  KSh5,300
List Price:  E£6,600.72
You save:  E£1,886.05
List Price:  ብር8,069.29
You save:  ብር2,305.67
List Price:  Kz119,054.91
You save:  Kz34,018.12
List Price:  CLP$127,353.10
You save:  CLP$36,389.20
List Price:  CN¥995.31
You save:  CN¥284.39
List Price:  RD$8,255.53
You save:  RD$2,358.89
List Price:  DA18,823.05
You save:  DA5,378.39
List Price:  FJ$317.44
You save:  FJ$90.70
List Price:  Q1,091.28
You save:  Q311.81
List Price:  GY$29,398.65
You save:  GY$8,400.21
ISK kr13,807.61
List Price:  ISK kr19,331.21
You save:  ISK kr5,523.60
List Price:  DH1,396.96
You save:  DH399.16
List Price:  L2,483.42
You save:  L709.60
List Price:  ден7,975.18
You save:  ден2,278.78
List Price:  MOP$1,129.51
You save:  MOP$322.74
List Price:  N$2,572.10
You save:  N$734.94
List Price:  C$5,169.58
You save:  C$1,477.13
List Price:  रु18,701.28
You save:  रु5,343.60
List Price:  S/524.20
You save:  S/149.78
List Price:  K545.83
You save:  K155.96
List Price:  SAR525.05
You save:  SAR150.02
List Price:  ZK3,718.29
You save:  ZK1,062.44
List Price:  L642.23
You save:  L183.50
List Price:  Kč3,193.44
You save:  Kč912.47
List Price:  Ft49,741.85
You save:  Ft14,212.97
SEK kr1,070.73
List Price:  SEK kr1,499.06
You save:  SEK kr428.33
List Price:  ARS$124,628.98
You save:  ARS$35,610.82
List Price:  Bs970.47
You save:  Bs277.29
List Price:  COP$536,293.92
You save:  COP$153,237.77
List Price:  ₡72,014.29
You save:  ₡20,576.98
List Price:  L3,470.95
You save:  L991.77
List Price:  ₲1,056,195.26
You save:  ₲301,791.63
List Price:  $U5,387.22
You save:  $U1,539.31
List Price:  zł550.04
You save:  zł157.16
Already have an account? Log In


All right, everyone. Welcome back. Now we talked about hierarchies. We talked about Cisco three layer models. We talked about a little bit about redundancy, the Cisco three layer. Well, we said Cisco dealer model ready.

All right, but we're gonna compare these two. And you you seen this already. But let's take a closer look at it. This is something that we want is a Cisco three layer model. This is your core, right? That's going to communicate to other campuses or their buildings, we don't want to overly burden this particular switch, because this switch is going to its main purposes, high speed data, okay.

Then you have your distribution switch, which communicates with your access switch in order to send money or what have you have access lists on here, you have VLANs on there, you can have all policies, whatever. Okay, these are layer three switches, you have routing policies on there as well filters. And then these are your access switches. These are the ones that go to your end devices. So this is a modular design, which is really a hierarchy, a hierarchy. But we've said it that that type of model does not give you what redundancy.

And this is the redundancy design. As you can see here what we did, we added another distribution switch in case that distribution switch or whichever switch goes down, you still have another switch that people can travel can flow through because you can see that you are connected to both switches. So this switch goes down, knows we're no worries, we're still both of these switches, access switches are still connected to this one distribution switch which goes to the core. And if you look at the core, we have a little triangle here, okay? That's going to say hey, you know, if a woman goes down, you can still send information to the other. But Cisco does recommend Cisco does recommend, they should have at least two core switches, two core switches, two distribution switches harmony access, which is unique, obviously.

And you know, this can be scaled to meet the level of your enterprise, right? They have a course which we're building a ton of distribution switches, and a whole lot of access switches, it all depends on the size of your particular network, the design will change. And you know, all that, but remember, right at the beginning, when I say that you should base your network design on the flow of traffic on the flow of traffic. Oh, come on. Let's use some common sense here. Okay, let's use some common sense.

When you're doing a network design. Yeah, look at what type of company this is. You're not going to develop something like this, which because this is beautiful is lovely. You know, you have two cores, three distribution switches, you know, a whole bunch of access switches. Everybody has redundancy. You're not going to have redundancy.

You're not going to have let's say from your end devices, you're not going to put two Nic cards in They're just have two different connections, you're not going to do that. Okay, so that costs money. So you got to get your budget first and see if the people that are in charge, the higher ups, the powers that be, will let you say, Okay, well, we'll go ahead, spend the money. No, no, it's no issue. Go ahead. Here's my black card.

Go at it. Okay. So that's never going to be the case. I you always got to take a look at that network that you're doing and say, Okay, what can I get away with? All right, while still giving a the most efficient network and secure network that I can I so here's that simple hierarchical designs, which you just saw, does not address the issue of redundancy. It doesn't.

It just doesn't and we need redundancy that we have not only we have physical redundancy, but we have our redundancy protocols, it is important to have redundancy. If one thing fails, then let him go another way you have to do that. Now that doesn't address the issue of scalability large additions to a network, it really doesn't because if you only have one core switch, my God, what do you do? Are you gonna over burn that switch with? Alright, let's just connect this just keep putting interfaces on that. on that one switch.

Just connect everybody on there. We don't care. We have a million users just put them on there. Come on, can't do that. That's why we build you expressways only because the flow of traffic is just too much for that four lane highway. Now we need a six lane highway only to make another highway that's going to go somewhere else so we need to make another exit somewhere else.

Because we have only an exit to say on this street or this avenue. Yeah, the street and then three miles or five miles down the road. You have another exit Well, maybe you want to have it every two miles you want to have an exit to release to get rid of some of that traffic because Mary that that day data, all that car doesn't need to go all the way up there. It needs to get down here somewhere. So you need to fix that it is the same thing here. Okay, you need to create you got to look at the flow of traffic.

Okay, where is the problem worse to slow down? How can I fix that? And that's where you start putting in your devices. All right, it fully done a hardcore network design shop no more than one core. What a fully redundant hardcore network design should have Oh should have sorry, should have more than one core and one distribution switch. We said we added distribution we just didn't add another core.

Alright, but you want to add another core, you could divide enterprise campus networks into the following basic elements. Here we go. Switch block. Definition as per the book, a group of access layer switches. Together with our distribution switches, this is also called access distribution block. So switch locks are just that.

Okay? They're access layer switches that are connected to distribution switches. Okay, there's just one big block. Okay? A segment, if you will, the core, the campus network backbone that connects all switch blocks. So the core, the core, I mean, the core, the name says the core is going to interconnect all the other switches together all the other segments together, right?

It's going to so you can send information back and forth. So it's, it makes sense to say Listen, you need more than one core. You may need 345 Who knows? Okay, you may or cor per building because there's so many notes because it's not only computers, we're talking about printers, plotters, projectors, okay, it's also about wireless itself. So you may need more than one core. All right, more than two more than three.

Got a look at what you're working with this is just to give you an idea and a concept of what we're talking about. But you as an engineer need to know, what exactly are you working with? One of the things that I learned, go off on a tangent just a little bit when you're creating a wireless network, is that you do not only a radio site survey, where you go, and you see what frequencies may cause what interference with your wireless, but you also do a physical site survey. Okay, where can I place the devices? All right, the radio site survey will tell you hey, this is the end. This is how far I can go with this particular router.

We have dead zones over here that we need to cover the physical side service as well listen to friggin steel ceiling here. You know, you got steel walls. You can't put an antenna here because, you know, it may fall and hurt somebody, or whatever the case may be, you need to do that type of survey. You have to make sure you know or to know where you put the equipment. The same thing goes with this. Okay, the same thing goes with this.

Okay, switch box, a group of access layer switches together with their distribution switches. Okay, well, how many? How many? You got to know what you're working on. You just can't say, Oh, yeah, I'll say the job. How many or how many computers you have?

Oh, you have too many computers. Yeah, we're gonna create, Oh, you don't need like 10 switches, and we'll do like four distribution switches. And we'll just call three cores in there. It's gonna cost you about I don't know $120,000 to create this network. It's like what? Wait a minute, wait a minute.

It's ridiculous. You got to know what you're working with. You got to go out there. You got to go on site and look to see what's going on. Okay. Now sizing a switch block.

Certain factors that should be taken into consideration when doing a switch block, what we're just talking about right now. range of available. So just make the switch block size, very flexible and the axillary switches selection is usually based on port density, meaning the number of connected users. So if you have five users, you really don't have anything to worry about. If you have 10 users, 20 users. All depends what you're doing.

If you go to an architect, architectural company, what have you, or a surveying company? Believe it or not, I've done I've done serving in my life. Okay, they use what we call a program. I always forget I remember sometime later, the one that draw out Okay, there we go. AutoCAD. AutoCAD files a huge, huge right architects work with this thing.

So I don't know that's something new or not at all haven't been up to date with it. So but I don't care does create a very large file. So if you're sending this information through the network woman You need to see the port density, right? How much is going through that port? Alright, so you can say okay, I need to have a gigabit switch. I need to put this you know, gigabit ports all this good stuff, but where am I placing this?

Okay, what distribution what distribution ah switch among other things Am I going to segment the people that are doing AutoCAD for the people that are just doing spreadsheets okay? Or the villagers working with you know, I don't know the receptionist I just training okay, because the AutoCAD is to the plotter and they're printing to a plotter so that plotter will be part of that block. Right. And that's what they're talking about here is a segmentation really fancy name, but a segmentation. Distribution layers must be size according to the number of access layer switches obviously. Obviously, the nearly no distribution layers must be sized, according to the number of access layer switches, if you have 200 access layer switches, meaning switches are connected to end devices 100 or 200.

What are you gonna get three distribution switches to? No, no, you're gonna get more than that. All right. So you need to decide But again, it's all based on, well, are they playing bingo, then we really don't. They really don't even matter. It doesn't even really matter.

They're just playing bingo and it's online or what have you. It doesn't matter. Okay, or in the local network. I can tell you right now that my students, I tell them okay, you want to play Call of Duty. Now we're in the classroom. So they actually did.

They dad they put the switches on the right that the routers on the rack. They punch down everything to the patch panel. They could put the wires to the computers they put in their IPS. We were able to communicate with each other. Let's play. Okay, but we need to go crazy with two cores, two distributions.

Now you didn't need to do that. You need to do none of that. Okay? One switch was one up unless you want to segment everything. Yeah, you guys play on your own playing around and that's it. Okay.

So the trick was it as switches that are aggregated or brought into a distribution device. What do you look at traffic types and patterns flow traffic amounts of layer three switches capacity at the distribution layer. Why layer three because are we routing? Are we routing this information? Is this now going up? Higher to the OSI, the OSI model?

Are we dealing with a routing protocol? Because if it's OSPF now we got to think about the hellos. Okay, we're thinking about the essays y'all assume right the updates the acknowledgments or the database LSP databases are being formed the SPF algorithm, the shortest path for us right to calculate all these different things, the updates, total number of users connected, we already talked about that. And if there's any geographic boundaries, or subnets, or VLANs is the subnet or the VLAN. is a really like, somewhere off across a street somewhere. You need to take that into consideration because what type of cabling is going through there?

Are we using optical fiber or are we using coaxial to get across? What is it that we're using? are using METRO IE if it is, is it going through a course which was what had to go to the course which, okay, everyone's somewhere else. So all these things need to be taken into consideration. But again, fancy fancy things, for things that we've already talked about. And the CCNA.

We know we need to segment we know we need to create VLANs you know, you cannot burden your VLANs then they're your Not using a VLAN for what it's for, yes, you're logically segmenting it, but now you're burning it. By putting 2000 generally no one 2000 users should be placed within a single block within a single segment. So if the whole point is for you to say, Okay, I'm going to segment my network into different blocks or segments or whatever you want to call it. Okay. I'm going to go ahead, and I'll put 2000 people on there. No, because you're creating the same issue.

Because again, Unless Unless that block, you further subdivided into VLANs. And in each VLAN, you'll have I don't know 500 users create for VLANs 500 users. Okay? All right. Now we can maybe work with it, but we need to see the amount of traffic that leaves it because remember the rule 80% of your resources need to be local to your segment. Okay, so here this term no more than 2000 users should be placed within a single block.

It's like me telling you, Hey, don't put 2000 people on your VLAN. I mean, come on our school that I one of the schools that I've been with, they had like 700 students at one time, right? Ah, it was a technical school. That's 100 students. But they're all been on one. So yeah, so 100 people on one VLAN.

Now, you know, I'm not telling every students like this. This is just human nature. What do you do? When you first open up your computer? You don't go straight to work. You check your email, you check your Facebook.

Okay, you check. I know some of you go to Snapchat, start taking pictures of yourselves with doggy faces and all sorts of stuff, okay. And you start putting that on the network. Alright, so you're already creating traffic without doing anything, or the IT guy might be late. from doing his imaging the night before, so he decided to do it now. So now everybody's getting a fight because an image is humongous, humongous, okay?

All right. So instead of switch box sizes should be based primarily on the following the types of traffic behavior for the traffic types and behavior, and the size and number of common work groups. We're repeating ourselves, we're saying the same thing over and over and over again, we know this, we need to limit this we need to segment physically and logically. And we need to limit the amount of people that are going to be in that particular segment, period. That's it. I know who they are.

You create these segments, these blocks, right. And each block we have a core that's connected to the main core. That's going to take you over the place. All right. All right now, this is Cisco's new thing. A collapse core or at least knew the certification was in the CCNA, which, in the CCNA, it says something completely different than to this.

But okay. I'll say to this, a collapse core block is one in which the hierarchy core layer is collapsed into the distribution layer. Both distribution and core functions are provided with the same switch devices. And here it is. Okay, so you have the core right here. Here's the core.

Okay. And now we have access and distribution. Right here. You see you have your redundancy. You have three access layers, right? They're all multi layer switches.

All right. So yeah, they can do you know, layer two and layer three. All right. They have three and a two distribution. They both have redundancy to the district. All Access switches have redundancy to the distribution switches.

We have the distribution switches, having redundancy to the other distribution switches. I save another They're building. Alright. And then we have the same thing with the access layers. And then the core is connected to both distribution switches. So let's say this link goes down, this individual that's sitting right up here, can, okay, I can go this way.

But I can go this way, this way and then go out that way. Now, obviously, we don't see the connection of it, but that will be the idea. Okay, we're distributions which would then look at a table a forwarding table, and say, okay, you need to go this way. Okay, if you need to go to another, you know, campus or what have you, I but here's what they mean by collapse core at the CCNP level, class core, and I make that very clearly because at the CCNA level, it was something different. They talked about distribution and access layer being together and the core being separate. Now here, they're putting a twist on it, and they're saying that a collapse core block is one in which the hierarchy core layer the core layer is collapsed into the distribution layer, and that the distribution and core functions are provided with the same switch devices.

Okay, so that was another definition in the CCNA. Okay, based on the book, but you follow what it says for that certification, you follow what it says for that certification. So this is what your collapse design is. And that's it. That's it. Basically, we repeat ourselves a bunch of times.

Segment logically and physically have redundancy in your network. Do not overpopulate your switch blocks, whether it be in the distribution or access layer or core layer. Okay. You do not want to burn it. Like I said, there are no more than 2000 users. That's a lot.

Okay. So this is what we're talking about. This is what we're talking about and this will be doing. There is a labs I'll show you okay in there. For you to go ahead and get a better understanding of what they're talking about seeing the next one

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.