Types of Network Topologies

Why topology matters

Obviously we are going to want to interconnect these different devices so we can communicate across our networks, but there are different ways we can set that up.

Here is a network that I commonly use for a lot of my demos. We have a LAN over here and a LAN over here, and then it is interconnected by these WAN devices. These are layer 3 routers right here that are going to route our traffic from this network over to this network. But there are different ways that we can set up these WAN connections.

Some of the ways that we can interconnect this is maybe with a point-to-point, or a bus, or a star, a ring, a tree or a mesh. We even have partial meshes and hybrids. So let's take a look at what these different terms mean and how we can interconnect our devices.

Just from an overview, here are some examples of these. Here is the point-to-point, where we just have two devices that are communicating. Or we have a bus, where they are all communicating to each other at the same time. Or we have a tree, which is some sort of hierarchy. There is a star, there is a ring, and there is a mesh.

Point-to-point

Point-to-point is perhaps the simplest out of here. We just have two devices and they talk directly to each other. So this would be a point-to-point connection. We even have things like point-to-point protocols that allow us to set up this connection between two routers or between two devices. So there are different ways we can set things up using point-to-point.

Bus

This is what a bus looks like. A bus is when all the nodes of a network are just interconnected. This was something that was very common when we started networking, but we have moved away from it from a networking perspective.

It is still really commonly used. For instance, USB is universal serial bus. It is a bus. And there are also buses on our motherboards. So it is still really common to use buses, although from a networking perspective it is just not as common to interconnect our devices using a bus, because of the drawbacks of it.

Star, or hub and spoke

This is a star topology, because it looks like a star. We also call this a hub and spoke, because in the center here we have a hub, and from here they branch out and we have these different spokes that connect these nodes right here. So star, or hub and spoke, is really common for our layer 2 networks and sometimes our layer 3 networks. This is a really common topology that we use a lot nowadays.

Ring

This topology is a ring, and it is because it looks like a ring. Of course, the more nodes that you get here, the more and more it will look like a ring.

Ring was used a lot back when we started networking. For instance, token ring was used quite a bit. We do not really use token ring anymore, but there are other types of rings. We still use this from a layer 2 and layer 3 perspective for self-healing. For instance, a SONET ring we still will use nowadays. So it is still a common topology, although we do not use it as much in our local businesses. It is more that your internet service providers use these rings.

Tree

Here is a tree. A tree, you can see, if you flipped it upside down, looks like the branches of the tree. Or you can think of it as the roots of the tree here. This is where we have some sort of hierarchy, where there is maybe a root switch here, and then from there we branch out and we have nodes or other switches here, and then data would flow up to the top here.

Now, this does not really offer a lot of self-healing. So we use this to a certain degree quite often nowadays, but we would probably have some other kind of interconnectivity between these devices for some sort of healing. And that actually works more towards being a mesh.

Mesh and partial mesh

Here is an example of a mesh. The idea behind it, or one of its real strengths, is that it is all interconnected. We see that each of these nodes is connected to every other node. Now, this could be problematic if we have hundreds of nodes. That gets a little bit too complicated, too many connections there. So maybe we want to limit how many nodes we are connected to. But often mesh is a great way that we have high availability, because if one of these connections goes down, there is still a lot of connectivity here and it really does not affect the network all that much.

Now, if it does become problematic to have too many nodes all connected together with each other, then maybe what we do is we roll out a partial mesh. This gives us some of the advantages of the self-healing and availability that a full mesh would have, but also does not compromise some of the things like too many connections. So here we can see some of these nodes are interconnected, but we have dropped a couple of the connections there. So we still have a lot of availability, just with a few less connections.

If we have a partial mesh, then we just call this other one a full mesh. When all devices are connected to all other devices, now we have a full mesh.

Hybrids

More often than not we are not just using one type of topology. What we are doing is intermixing these different topologies. So here we can see that there is a tree on one part of our network, and then we have a full mesh on another part of our network. A lot of times this would be like our layer 3: we want a lot of high availability, so there would be a full mesh with that. Versus over here, maybe this is our switching on one of our LANs. So it can look different in how we set it up, and we just call this a hybrid.

Here is an example of that network topology map that we had at the beginning. Can you see what kind of network that we have here? From a layer 3 perspective, we have a full mesh. We can see that they are interconnected; every single router is interconnected with other routers on here. From a layer 2 perspective, we see that we have a little different setup here. We just have a hub and spoke.