With a peer-to-peer network, there's not a server or a client. Instead, every device is a server and a client, so everyone is communicating to each other to be able to gain access to the data that they might need. This is a very easy way to deploy an application, because you simply install it on everyone's computer, and they all talk amongst each other to be able to use that particular app. You don't need an additional server. You don't have to purchase and maintain that server operating system. Everything runs on everyone's existing systems, so the cost is relatively low to implement. One of the significant disadvantages of peer-to-peer, however, is that it's very difficult to administer this type of system. The entire application is distributed. All of the authentication process is also distributed, so it's very difficult to be sure that a peer-to-peer application will remain secure. With client-server communication, the responsibilities are split. You have individual clients, and then you have a centralized server, and when clients need to access data, everyone will access the same server. As you can see by this diagram, clients are not communicating to each other directly, or even through the server. All the clients talk to the server, and the server talks to the clients. This gives you very good performance, because you have a direct connection to the server that you can manage and maintain. This also allows for easy administration, because you can configure the application and the authentication for that app at the central point where the server is. As you can tell with this diagram, though, there's additional hardware and software that you'll need. You'll need a server, and you'll need somewhere to store it and maintain it, and that has an additional cost associated with it. This also adds a bit of complexity, because you have a centralized server that needs to be administered. That's something you don't have on a peer-to-peer system. If you're working inside of an office, or you're working from home, that network that's in your office or in your home is a local area network. It's local because the network that you're using is usually within the same building that you happen to be in. This could also be the network that's within a single building, for example, between different floors. That is a local area network, because all of those resources are local to us in that building. Generally speaking, a local area network will have a relatively fast throughput. We're using 802.11 wireless networks and ethernet connectivity to be able to have high-speed communication. If you find that you're using slower connections, or it's a link that doesn't run at these higher speeds, then you're probably not communicating over a local area network. Many companies will have buildings or parts of their organization that are located in different parts of a city to be able to communicate to those buildings, we need to use a metropolitan area network, or a MAN. This is obviously a much larger network than something that's in our single building, and it's usually expanding out to a large, citywide area. Usually, you're working with a third-party provider in that city to provide this service. Occasionally, you'll see this referred to as metro ethernet for this metropolitan area network. It's very common to see governments use metropolitan area networks, because they're obviously very geographically dispersed throughout the metropolitan area, and they have right of way, so it's very easy for them to put fiber in the ground and connect up all of their remote locations. If you're going outside the scope of a metropolitan area network and communicating over a much larger distance, then you're probably using a wide area network, or a WAN. One characteristic that tends to be very common when you start expanding the distance of these networks is that the networks tend to become a little slower in their throughput. So wide area networks generally don't run at the same speeds as a local area network that you might have in your building. If you're connecting over an MPLS network or point-to-point serial WAN connection, then you're almost certainly communicating over a wide area network. These can be links that are connecting us with fiber that's in the ground, but we could also use satellite communication, which is certainly a wide area network, and those nonterrestrial links would allow us to connect locations that may be very far away. A specialized type of local area network is one that is wireless. This is the WLAN, or Wireless Local Area Network, and usually these are 802.11 networks. If you're within a building and you're walking from one side of the building to another and you're connected to your local network, then you're probably connecting over one of these wireless LANs. This is one that does have a limited geographical area. When you leave your building, you're probably not going to be able to communicate to this WLAN because of the short distances associated with 802.11. But if you're in a larger area, there are ways to extend this distance using multiple access points. So you could be in an entire downtown area or in multiple buildings and still maintain connectivity to that wireless LAN. If you have a Bluetooth headset, you're connecting to your computer with a Bluetooth keyboard, or using some other type of personal connection between you and the devices you use, then you're using a personal area network, or a PAN. If you're in your car and you connect to a speaker and a microphone or you play audio through your phone to the audio system of the car, then you're using a Bluetooth-based personal area network. We also consider that Bluetooth connection to a wireless headset to be a personal area network. And if you're using some newer workout equipment at your gym, you'll find that is connecting to your mobile phone using this personal area network. There's a middle ground between a local area network in a building and a metropolitan area network that might connect us to a site that is many miles away. We could have multiple buildings within our local area on a single campus. You certainly would see this with schools and universities, but this could also be the same for large organizations that may span multiple buildings in a particular area. This is a campus area network, or a CAN, and it's very common to have these buildings connected with conduit, making it very easy to pull fiber from one building to another and connect all of your users to each other. We would usually run some type of local area network technology so we would have high-speed ethernet between all of these buildings, so that we're able to communicate as if we were sitting all in the same building. This is usually a connection that's made on private LAN. You've put your own conduit into the ground, you've run your own fiber connection, so usually there's not a third party that needs to be involved. There wouldn't be a monthly payment to an internet service provider because you're the one that has built the conduit, you've run your own fiber, and you don't have to involve a third party. One of the reasons for adding all of these connections, especially on local and wide area networks is so that users can have access to their data usually this data is located in a central location and we're usually storing it on something like a network attached storage device, or a NAS. This NAS contains multiple drives and is able to store large amounts of data in this one single area. A network attached storage is usually a device where you can access these files at the file level. File-level access means that any changes you make to this file have to be done to the entire file that's located on this device. This means if you have a file that's a gigabyte in size, and you just need to change one byte within that file, you have to overwrite the entire gigabyte, because you're using a network attached storage. If you just wanted to change that single block of data that contained that information, then you would want to use a storage area network, or a SAN. To your computer, a SAN looks and feels a lot like the local drive that's on your system. So being able to make block-level changes to this information can be done very quickly and very efficiently when compared to something like a network attached storage device. Whether you're using a NAS or a SAN, you're probably going to be sending and receiving a lot of data to these devices. For that reason, you would probably have a dedicated network for the storage area networks or network attached storage devices, and you would probably connect these with the fastest network possible. Through the years, we've had a number of different wide area network technologies. We had frame relay and ATM, and as our needs changed and the way we use these networks changed, we needed a smarter way to communicate. So we created MPLS, Multiprotocol Label Switching. This is communication through the WAN, but it uses labels to be able to determine how we route and forward that traffic through the WAN. One advantage of MPLS is that we can use almost any type of connection for MPLS and we can put almost any type of data inside of an MPLS network. This means that we could send IP traffic, ethernet frames, or any other type of data we'd like through the MPLS network. This makes MPLS relatively easy to get up and running and have all of our data transferred between two locations. With IP routing, we're always concerned about the IP subnetting and the next hop that we're going to use. But with MPLS, we make forwarding decisions based on the label switching. This label switching is something that's added when we put information into the MPLS network. If we have data and we have information on one side of the network, and we would like that data to get to the other side of the MPLS network, we send that data into the first provider edge router that's closest to us. That edge router is going to insert a label into this data. We call that pushing the label onto the data. It's going to then know how to forward that traffic through the internals of the provider switch network, and when this data reaches the other side that label is popped off of this communication. And the remaining data is sent to the customer edge router. With MPLS, we have to create an initial configuration that defines where all of the different sites may be located and what labels may be used to switch data to those locations. Other types of wide area networks could create network connections dynamically, so that you're only connecting to those sites when you need to. This commonly uses MGRE, which is Multipoint Generic Router Encapsulation, and you'll usually see networks like a dynamic multipoint VPN, or DMVPN, used to send data across one of these MGRE networks. What's nice about these dynamic connections is they're only created when they're needed, and they're torn down when you don't have any information that you need to send. If you have a break in any of the connections, the VPN can even rebuild itself to be able to communicate between all of these different sites. All of this is considered to be a dynamic mesh because you're able to send data to whatever site you would like based on where that data needs to go. Normally, you would have a main office, and you would have remote sites that would connect to the main office over single links. And if you needed to communicate between remote sites, that remote site would have to first communicate to the main office, and then the main office would have to communicate down to that destination site. With a dynamic multipoint VPN, you can have these two sites recognize that they want to communicate to each other, they create their own link between the communication, and when that link is no longer needed, it's disconnected. If this site needs now to connect to a different remote site, it can then create a separate dynamic link to that site, and of course you can have that site create a dynamic link as needed. All of these dynamic multipoint links mean that you have a more efficient network. You're only using connectivity when you have to have it, and if you do need to send information between two remote sites, you can send that data directly rather than hopping through a main office. A newer type of wide area network is an SD wide area network, or SDWAN. The SD stands for Software Defined, which means that our wide area network can be defined however we would like to lay it out. This is especially useful if you have cloud-based applications and you need the network to understand where data needs to go at any particular time. We used to have a data center that was located in one central building, and we simply connected all of our remote sites to that one data center. Now that we have cloud based applications, we need to communicate to the cloud. And instead of having to hop through this data center to be able to access those cloud-based resources, we can have the wide area network recognize that we need to communicate directly to those resources rather than sending all of that data through one central data center. Here's our wide area networks before we had cloud-based services and before we had software defined wide area networks. We had our data center. Inside of our data center were all of our services, our email servers, our database servers, and all of our other data. All of these remote sites would need to connect to the data center to be able to gain access to those resources. With cloud-based services, we've taken our email services, we've moved them to the cloud. All of our servers are now at a cloud provider, and we might have our databases inside yet another cloud provider. Instead of having all these remote sites then communicate to this central data center, and then hop to all of these cloud-based services, we could have our wide area network automatically know where to send the data based on the type of data that it is. So if a remote site needs to access data that's in a cloud database, they can access that information directly from the cloud. If another site needs to grab an email message that's located in a cloud-based email service, they can communicate directly to that email service instead of hopping through a central data center. The determination of how this data is routed is in the wide area network itself. That software defined aspect of the SDWAN determines how data is rooted and sent based on the type of services that are required. This makes for a much more efficient network. This allows us to have instant access to our cloud-based resources instead of having to hop through a central point or build out a separate network to gain access to those resources.