Understanding traffic flows across topologies can help with the design and troubleshooting of our networks. In this video, you’ll learn about topologies such as mesh, bus, ring, star, and hybrid networks.
<< Previous: Network Cabling ToolsNext: Network Infrastructures >>
If you’re in the process of building out a new part of your network, then you need to understand exactly the advantages and the disadvantages of the network topologies that you may be choosing. And in this video, we’ll look at a very broad overview of some of the most popular network topologies that you’ll run into. These will help you a lot with not just planning, but once the network is in place, you’ll get an understanding of exactly where the signal of the network is flowing and where you might run into problems with maintaining uptime in certain parts of the network.
When we refer to a network that has a mesh, then it’s one that has one or more connections to multiple sites. This may be a fully meshed network where every site is connected to every other site. But generally, we have partially meshed networks so that we can have some sites with multiple connections to other sites, but perhaps not to every other location that we have.
This is obviously a good type of topology if you’re worried about uptime because you have redundancy. If you lose one link to a site, you can simply use the other link. And since we’re all meshed together, we can eventually get that data back to where it needs to go.
You might also be able to do load balancing. You may have one site at the top, and to be able to connect down to another you might send traffic over multiple links simultaneously, taking advantage of all of these redundant links that you’ve created. It’s very common to see these on Wide Area Networks since we have very little control about what’s happening with the network outside of our buildings. This way we can build multiple links to a single site, and if we lose one of those links, we still have the other as a redundant link because we’ve created this mesh.
An older topology style that you still might run into is a bus topology. This is one where you effectively have a single cable and everybody is connecting into that cable. It’s a single bus that everybody is using and communicating through.
This is a very simple network. It’s easy to install. Everybody connects to a single line, but it is prone to errors. If you lose one link in the middle of this, it will break this connection and now suddenly half of the network can’t see what’s happening on the other half of the network. Also, if one person was to inject any type of error signals onto this link, it is going to affect everybody on that bus.
We often think of ring topologies as being those older style of networks. When we think back to the older token ring networks for instance. And although we don’t see token ring networks around much any longer, we still use ring topologies for our Wide Area Networks. A good example of this would be something like SONET, where you can build a redundant ring SONET network in a geographical area.
The SONET networks work by sending exactly the same information down both sides of the ring. So if you do have a break anywhere along the way, that information will be able to make its way by going the other direction on the ring.
If you’re connecting to an ethernet network right now, you’re probably connected to a star topology. A star topology is one where all devices are connected through a device in the middle. Gives it this look of a star because you have this central device. And on an ethernet network, this might be a switch whether you’re using a connection in your home or home office, or whether you’re in the largest topology in the biggest companies in the world, they’re all connecting back to some central switch and have this star topology in place.
A hybrid network is when we’re combining a lot of these topology types into one single view. Here’s a good example of having multiple star topologies. And then we’re connecting all of these sites together with a ring. We’ve taken different kinds of topologies, brought them together into something that is now a hybrid of multiple types.
A point-to-point topology is exactly the way it sounds. You have a device on one end, you have a device on the other end, and they’re connected to each other with nothing in the middle. We see this often with Wide Area Network links that are a little bit older like our T-1 or E-1 connections. These are point-to-point Wide Area Network connections.
You may be connecting buildings together. There may be a single strand of fiber or copper between the buildings. You might also consider that a point-to-point connection. The important thing is that one device is connected to the other device, and there’s really nothing in the middle. There’s a single point-to-point between them.
Another popular networking topology is the point-to-multipoint. If you’re on a wireless network right now, it’s probably something like a point-to-multipoint network. If you’re using something like this wireless network, a point-to-multipoint doesn’t necessarily imply that everybody can talk to everyone else. It’s really that everybody is talking to one central device. Most wireless networks will allow you to talk to other devices on the network, but some wireless networks are configured to not allow these wireless devices to communicate with each other. They can only talk to the center with this point-to-multipoint topology.
Let’s move away from a pure network topology conversation and talk about more of the way that applications are architected. In this particular diagram, we’re showing a client-server application. This is where the clients are talking to a central server. Your client might be a laptop, or a tablet, or a desktop workstation.
And to be able to use this application, everybody talks back to a central server to be able to perform these functions. The clients do not talk to each other, so there’s no client-to-client communication. And one of the advantages of doing it this way is that you have performance advantages because your server’s really going to determine the performance of the application. And it’s easy to administer because you go to one place on the server to be able to administer this app.
This also means that you’re going to have a little bit higher cost because this server needs to be strong enough to support all of these separate clients that are using this application. And there’s a bit more complexity because you need separate clients that are going to be installed or running on these end stations. And they all have to communicate across a single network to a single set of servers.
Another popular application type is the peer-to-peer application. In this case, there’s no central server at all. The different clients are all communicating between each other to perform whatever steps they need for this application. Every device is able to talk to every other device on a peer-to-peer network.
This is obviously easy to deploy since everybody’s probably on the network already, and they can communicate to each other. And there’s a lower cost because you don’t have a central server that has to be built and administered. The disadvantages here are that it’s much more difficult to administer. There’s no central point for the application. So now you are effectively administering this application on every single desktop.
You also have to be very careful about security since every device becomes both a client and a server in a peer-to-peer network. You have to make sure everybody is up to date with all of the security pieces they need and that you’re able to keep that data as safe as possible.