Free CCNA | LAN Architectures | Day 52 | CCNA 200-301 Complete Course

Welcome to Jeremy's IT Lab. This is a

complete

and

free

course

for CCNA. If you like these videos, subscribe to follow the series. Also, like, leave a comment, and share the video to help spread the word about this

free

video series. Thanks for your help. In this video we will see LAN

architectures

. By

architectures

I mean how devices like routers, switches, PCs, servers, etc. are connected. to create networks. These LAN architectures are exam topics 1.2.a, b, c and e. 2 tier, 3 tier, spine and small office/home office. The other topics in 1.2, 1.2.d which is WAN and 1.2.f which is on-premises and cloud, will be covered in separate videos.

This is what we will cover in this video. Actually, they are exactly the exam topics I just pointed out. First, we will cover tier 2 and tier 3 LAN architectures. These are common traditional LAN designs used in enterprise networks. We will then cover the spine-leaf architecture, which is a common design in data center environments. Finally, I will briefly cover SOHO, small office/home office networks. Small networks like this differ from others in that they typically have a single network device that acts as a router, switch, firewall, and wireless access point. Watch to the end of the video for an additional ExSim for CCNA practice question from Boson Software, my recommended CCNA practice exams.

Before we get into the topics of this video, let me introduce you to why this section of the

course

is important. You have studied various networking technologies up to this point in the course, such as routing and switching, STP, EtherChannel, OSPF, FHRP, switch security functions, etc. So, now let's look at some basics of network design and architecture. Network design is a very deep topic and to be a good network designer, a good network architect, requires a deep understanding of the general principles as well as the needs of the specific client. Although there are standard best practices for network design, there are very few universal right answers.

The answer to most general network design questions is "it depends," because each network's requirements are different. Now, in the early stages of your networking career you probably won't be asked to design networks yourself. This is usually left to more experienced engineers. However, to understand the networks you will be setting up and troubleshooting, it is important to know some basic network design and architecture concepts. So in this video and the next ones we'll cover some basics. And now, before I introduce the two- and three-tier LAN designs that are the main topics of this video, I want to briefly introduce some common terminologies that you may hear and might use throughout this video.

These are general terms, not only for LAN design but for connections between devices in general. The first is called a "star" topology. When multiple devices connect to a central device, we can draw them in a star shape as shown below, which is why this is often called a star topology. In the diagram below, all the PCs are connected to a central switch and I have placed them in a star shape around the switch. Note that in network diagrams devices may not be drawn like this, in a star shape. But if many devices are connected to a central device, we can call it a star topology regardless of how the diagram is drawn.

The next term is full mesh, which is when each device is connected to each other, forming a mesh. Here is an example. 6 routers, and each router is connected to each other. This is called full mesh topology. The last one I want to introduce is partial mesh, which is when some devices are connected to each other, but not all of them. Here is an example with four switches. Notice that the two top switches are connected to each other, but the two bottom switches are not connected directly. This is a partial mesh topology. Well, those are the three general terms I wanted to present.

You'll see examples of them throughout this video and the next, so be sure to remember them. Now let's get into the basics of campus LAN design. By campus LAN I mean a LAN of devices in a building, or perhaps multiple buildings, that are close to each other. Remember, LAN stands for Local Area Network. We will first look at the two-tier LAN design, which consists of two hierarchical layers: the access layer and the distribution layer. The two-tier design is also called a “collapsed core” design because it omits a layer found in the three-tier design, the core layer.

Or more accurately, it combines the two into a single layer. So on the next slide I'll show you an example of a two-tier design, but first let me summarize each layer. First, the access layer. This is the layer to which end hosts connect, for example PCs, printers, security cameras, etc. Typically, they will connect to an access layer switch. Therefore, access layer switches typically have many ports to which end hosts can connect. QoS marking is normally done here. This is the layer that end hosts connect to, so we typically mark traffic here because it is good practice to mark traffic as early as possible on the network.

Security services such as port security, DAI, etc. are also usually carried out here. Additionally, the switch ports can be enabled for PoE and Power over Ethernet if there are wireless access points or IP phones on the LAN. That's the access layer, the switches that the end hosts connect to. Then there is the distribution layer. These switches aggregate connections from access layer switches. So depending on how many end hosts there are, your network could have many access layer switches. Those connections are then typically added to a redundant pair of two distribution layer switches, as you'll see on the next slide.

Typically, this is the boundary between Layer 2 and Layer 3 in the network. Therefore, distribution layer switches run Layer 3 protocols, such as OSPF, and Layer 2 protocols, such as spanning tree. This is not always the case, but typically the connections from the access layer switches to the distribution layer switches are Layer 2 connections, and then end hosts use the SVIs on the distribution layer switches as their default gateways. The distribution layer is used to connect to services like Internet, WAN, etc. Well, let's look at an example. Below is a simple two-tier campus LAN. The two access layer switches, A1 and A2, each have a few different end hosts connected and each also has a wireless access point connected.

D1 and D2 are the distribution layer switches and each is connected to A1 and A2. This provides redundancy by having additional connections in case one fails. However, the connections between the access layer and the distribution layer are Layer 2, and that means loops can occur. Therefore, the spanning tree protocol disables some links to prevent loops. Now, D1 and D2 are multilayer switches, and end hosts use the SVIs on D1 and D2 as default gateways. Therefore, they will likely use a first-hop redundancy protocol such as HSRP or VRRP to provide a redundant virtual IP address for hosts to use as the default gateway.

As I said before, the distribution layer is used to connect to services like the Internet or the WAN, as well as other parts of the LAN. So here's the Internet connection, two redundant connections to two different Internet service providers. And maybe there is another set of distribution and access switches here, with many servers connected to the access layer switches. Note that each distribution layer switch is connected to each other. By the way, in a two-tier collapsed core design like this, the distribution layer is sometimes called the core distribution layer, because it has to serve the purpose of both layers.

We'll look at the core layer later. These connections between distribution switches are Layer 3, no spanning tree is running here. Routing information can be shared over OSPF, for example. This is an example of a two-tier network architecture. The end hosts connect to the access layer which does things like QoS marking, security features like DAI, etc. Distribution layer switches aggregate those connections and then connect further to services like the Internet and other parts of the LAN. Now let me relate this to the common terminologies I mentioned at the beginning. Although I haven't drawn the diagram as a star, at the access layer we have many devices, all connected to a central device, each access switch.

So these are four small star topologies. If I put more end hosts on the diagram and placed them around each access switch, it would look more like a star, but I think you get the point. And here we have a couple of partial mesh topologies between the access layers and the distribution layers. Note that the distribution switches are connected to each other and to each access switch, but the access switches are not connected directly. So these are two partial meshes. And finally, between the four distribution layer switches we have a

complete

mesh. Each distribution switch is connected to each other, forming a complete mesh.

You will see elements of full mesh, partial mesh, and star topologies in all types of network designs and often combinations of them, sometimes called hybrid topology, so once again be sure to remember these terms. Now, if the network grows, we may have many distribution layers connected to different parts of the LAN. Do you see anything wrong with this? In large LANs with many distribution layer switches, for example, in separate buildings on a campus, the number of connections required between distribution layer switches grows rapidly. This makes it much more difficult and complicated to scale the network, make it bigger.

So to help scale large LANs, you can add a core layer. By the way, Cisco recommends adding a core layer if there are more than three distribution layers at a single location. In this case there are 6, so we should add a central layer. Here's what it looks like with a center layer added. Each distribution layer connects to the core layer, without requiring a full mesh of direct connections between distribution layer switches. These core layer switches are a very powerful and fast pair of switches. So let's take a look at how the core layer works now. When we add a core layer, we now have a three-tier LAN design, access, distribution, and core.

Let's talk about that core layer. Connects separate distribution layers in large LAN networks. The goal of this layer is speed; You may hear the term “rapid transportation” as a core layer objective. CPU-intensive operations such as security functions, QoS marking and classification, etc. should be avoided at this layer. We only want fast forwarding of packages, nothing extra. And all the connections are Layer 3. We definitely don't want a spanning tree in the Core Layer. The core layer must maintain connectivity across the LAN even if devices fail. As it is the backbone of the LAN, redundancy of devices and connections is very important.

Let's take a look at the above topology with a core layer added. This is what the network could look like with the addition of a core layer. This time, the core switches connect to the Internet routers. And if we have additional layers of distribution and access, where will they connect? As you know, they will also connect to the central switches. This is an example of a three-tier campus LAN. For smaller LANs, two levels are sufficient, but larger LANs will probably have three levels like this. Okay, let's go over those levels one more time and then move on.

Here are those attributes of each layer again. For the access layer, remember that it is the layer that end hosts connect to and often performs services like QoS marking, port security, DAI, DHCP snooping, etc. For the distribution layer, remember that it is about adding connections from the access layer and serves as a boundary between the Layer 2 network and the Layer 3 network. In a two-tier design, it connects to services such as the Internet and WAN , although in a three-tier network normally those connections will be to the central layer. Note that I added an additional note saying that it is sometimes called an aggregation layer.

Remember that connection aggregation is an essential function of this layer. Finally, the core layer is used in large LANs to connect distribution layers together, and the focus is on speed, so we avoid operations such as security and QoS marking and classification. All connections here are fromLayer 3. Well, that's it for the two- and three-tier campus LAN designs. Keep in mind what I said before: there are very few universal right answers in network design. Although these two- and three-tier LAN designs are common practice and recommended by Cisco, in the real world there are countless variations depending on business needs.

Let's move on to the topic of Spine-Leaf architecture. This photo I'm showing you was taken in a data center, and the reason I'm showing it to you is because the spine leaf architecture was designed for data centers. I currently work for a large data center provider, so I spend a lot of time in environments like this. So what exactly is a data center? Data centers are dedicated spaces or buildings used to store computer systems, such as servers and network devices. They typically contain rack rooms, as in the photo above, and servers and network devices are mounted in the racks.

Traditional data center designs used a three-tier architecture, access, distribution, and core, as we just covered. This traditional layout worked well when most of the traffic in the data center was north-south. What does that mean? Well, if we look at this diagram, north-south refers to traffic from the access layer to the distribution, the core and then perhaps to the Internet or back south to the other distribution and access layers. These other sections of the LAN are drawn to the side so that the traffic is not facing north-south when drawing the arrows in this diagram, but of course it is also north-south traffic.

This is in contrast to east-west traffic, which, for example, traffics between these servers in the same part of the network so the traffic does not go out to other parts of the LAN or to the Internet. With virtual server priority, which is a topic I'll cover soon in the course, by the way, applications are often deployed in a distributed manner, across multiple physical servers, which increases the amount of East-West traffic in the hub. of data. . If you're not sure what a virtual server is or you don't know what virtualization is, like I said, I'll cover that soon in another video.

Anyway, the point is that with this increase in east-west traffic, the traditional three-tier architecture created bandwidth bottlenecks as well as variability in latency from server to server depending on the path being taken. take the traffic. To solve this, spinal leaf architecture, also called Clos architecture after the name of one of the designers, has become prominent in data centers. We'll see. This is what the spine leaf architecture looks like. It has two levels, but it is different from the traditional two-level design we saw before. There are column and leaf switches, and here are the basic rules about column and leaf architecture.

Each leaf switch is connected to each column switch. If you look at the diagram below, there are three column switches and each leaf switch has three uplinks, one for each column switch. Therefore, each column switch is also connected to each leaf switch. However, leaf switches do not connect to other leaf switches and column switches do not connect to other column switches. Finally, end hosts, for example servers, only connect to leaf switches. They are like the "access layer" of the spine sheet architecture. Well, those are the main rules of lomo leaf architecture. The path that the traffic follows is chosen randomly to balance the traffic load between the switches in the backbone.

And each server is separated by the same number of "hops", except those connected to the same leaf switch which of course have fewer hops, and this provides constant latency for east-west traffic, traffic between servers . For example, for this server on the left to reach this other server, the traffic goes through three switches. And then to get to this other server, you again have to go through three switches. So, that's the architecture of the spine. It's also very easy to scale, because if you need to add more servers, you can simply add another leaf switch and connect it to the existing main switches.

As with all CCNA topics, Spine-Leaf Architecture has a lot more depth than what I just mentioned, but make sure you know these basic features of Spine-Leaf for the CCNA exam and you'll be good to go. Now let's move on to the final topic. That last topic is SOHO networks. Small Office/Home Office, also called SOHO refers to the office of a small business, or a small home office with few devices. Now it is no longer necessary to use it as an office. If your home has a network connected to the Internet, it is considered a SOHO network.

SOHO networks do not have complex needs, so all network functions are usually provided by a single device often called a home router or wireless router. You don't need a dedicated device for each function. This device can serve as a router and connect your home network to the Internet. Also a switch, as there are usually some switch ports on the back to connect other devices. There are also usually simple firewall functions, usually configured to block all connections coming from the external network, but to allow Internet connections from internal devices. It also serves as a wireless access point, allowing you to connect your laptop or phone, for example, to the network via WiFi.

Additionally, in some cases it can serve as a modem for cable Internet, although sometimes the modem is a stand-alone device. Below is an example of a simple home router, although they come in all shapes and sizes. For very small networks, it is not necessary to have a separate device dedicated to routing, a device dedicated to switching, a device dedicated to providing wireless access, etc. This device can do it all. So while a business network would have them all as separate devices, the home router combines them all into one small device. Now, throughout this course we will focus on business networks.

Very small businesses or home offices do not have the need, or likely the resources, to hire a dedicated network engineer. Therefore, they often simply rent a wireless router from the ISP and that serves their purposes. Here's a review of what we covered in this video. First, traditional 2- and 3-tier LAN architecture. Make sure you know the three layers, access distribution and core, and their basic characteristics. I then introduced the spine-leaf architecture, which is common in modern data center environments. Again, make sure you know the basic rules of spine-leaf and its characteristics. Finally, we take a brief look at SOHO, Small Office/Home Office networks.

They are small networks that typically have a single network device that provides routing, switching, security, and wireless access. As always, look to the end of the quiz for an additional ExSim for CCNA practice question from Boson Software, my recommended CCNA practice exams. Well, let's go to question 1 of the questionnaire. Which layer typically serves as the boundary between Layer 2 and Layer 3 in a traditional 2- or 3-tier network? Pause the video now to select the best answer. The answer is B, distribution. Typically, connections from the access layer to the distribution layer are Layer 2 and run a spanning tree protocol.

The connections from the distribution to the core are Layer 3. So the distribution layer serves as the boundary between Layer 2 and Layer 3. Ok, let's move on to question 2. Which of the following would you NOT expect to find in the Core Layer? of a traditional 3-level LAN? Pause the video now to select the best answer. Well, the answer is B, STP, Spanning Tree Protocol. Because all connections at the core layer are layer 3, the spanning tree protocol should not run at the core layer. Ok, let's move on to question 3. At which layer would you expect to find PoE enabled switch ports in a traditional 3-tier LAN?

Pause the video now to select the best answer. Well, the answer is A, Access. Devices that use PoE, such as wireless access points, IP phones, and IP security cameras, connect to the access layer, so this is where you can often find PoE-enabled ports. Ok, let's move on to question 4. In a Spine-Leaf architecture, which of the following should not be connected to a Leaf switch? Pause the video now to select the best answer. The answer is B, a leaf switch. Endpoints, such as servers, can connect to leaf switches, and each leaf switch must connect to all spine switches, but leaf switches must not connect to each other.

Ok, let's move on to question 5. Which of the following features could be included in the device known as a wireless router? Pause the video now to select the best answer. Well, the answer is F. A wireless router, also known as a home router, is a multipurpose networking device that can provide routing, switching, wireless access, security features, etc. for a small network. This type of device is often used in SOHO, Small Office/Home Office networks. Well, that's all for the quiz. Now let's take a look at an additional question in ExSim for CCNA from Boson Software. There are complementary materials for this video.

There is a deck of cards to use with the 'Anki' software. There will also be a package tracking practice lab, so you can practice. In this case, although we don't cover any configuration in this video, I will take the opportunity to demonstrate an important concept for LANs regarding the interaction between the Spanning Tree Protocol and FHRPs such as HSRP. That will be in the next video. To get free flashcards and lab files for the course, register at the link in the video description. Before I end today's video, I want to thank the members of my JCNP level channel.

To join, click the "Join" button below the video. Thanks to Khoa, Dragos, Tanvir, Charlesetta, Gerrard, Tom, Samil, Scott, Martin, Tebogo, Anand, Pavel, Abraham, Serge, Njoku, Viktor, Suki, Kenneth, Seamus, Brandon, Marcel, Kone, Donald, Gustavo, Prakaash, Nasir, Erlison, Marko, Ed, John, Funnydart, Velvijaykum, Boson Software, Devin, Yonatan and Vance. Sorry if I mispronounced your name, but thank you very much for your support. By the way, this is the list of JCNP level members at the time of recording, July 3, 2021. If you recently registered and your name is not listed here, don't worry, you will appear in future videos. Thanks for watching.

Subscribe to the channel, like the video, leave a comment, and share the video with anyone else studying for CCNA. If you want to leave a tip, check out the links in the description. I'm also a verified Brave publisher and accept BAT, or Basic Attention Token, suggestions through the Brave browser. That is all for now.