Let’s talk about the different ways we can categorise computer networks. There’s actually a variety of different classifications we can use. Here we’re going to talk about three of them. We’re going to talk about host roles, we’re going to talk about geographic proximity, and then the signalling methods used. Let’s first talk about classifying networks according to the host roles. In other words, what do the hosts do in the network?
The first one we want to look at is called peer-to-peer. In a peer-to-peer network, network hosts don’t have a specific role that they play. In other words, hosts on a peer-to-peer network both provide network services, and hosts on a peer-to-peer network also consume network services. What exactly does that mean? In a peer-to-peer network, we have a variety of operating systems here, and they fulfil a variety of different roles. For instance, over here we may have a workstation that has a printer connected to it and that printer is shared on the network, allowing each of these different hosts to print to that printer. Over here, this host might have a huge hard drive installed and everyone is allowed to share that hard drive. If they have a huge file they need to save, they can put this file down here.
In this situation, you have hosts that both provide and consumer network services. These two hosts provide a new service. This one provides printing. This one provides storage. At the same time, these hosts also consume services. For instance, this host prints to this workstation. This workstation saves files to the hard drive in this workstation. In essence, they function both as a client and as a server at the same time.
There’s a lot of benefits to a peer-to-peer network. First of all, it’s very easy to implement. In other words, you could take a whole bunch of Windows XP workstations, for example, and create a peer-to-peer network. You can share printers, share storage. You don’t have to go to any trouble to configure this. All you have to do is share your resources. It’s very easy to implement. It’s also very inexpensive. In this case, with a Windows XP peer-to-peer network, you just instal the operating system. That’s it. There’s no special software to purchase and implement.
There are some drawbacks to a peer-to-peer network. You might be asking, “Well, it’s so easy to implement, so inexpensive, why don’t more companies implement it?” There are some key problems. First of all, a peer-to-peer network is not very scalable, meaning the bigger it gets, the harder it is to manage and the harder it is to keep running. Peer-to-peer networks are very, very difficult to support. That’s because they lack centralised control. In other words, there’s not one network administrator running the show. Every user at each workstation is like their own little network administrator.
Let’s take an example of how this is a problem. Let’s say that you have shared storage over here on this hard drive and this workstation belongs to Fred. Fred says, “You know, there’s going to be an awful lot of files on my hard drive. I think I’m going to erase a whole bunch of them. Let’s just erase them so I have more room on my hard disc.” What impact does that have on everybody else? Mary over here might have had her files on that hard drive, and he just wiped them out. She’s going, “My files!” No centralised control.
Further, let’s suppose Mary decides to get back at everybody because Fred over here deleted her files and she takes off the next day from work, leaves her office door locked and her computer off. Everybody needs to print. Can they print? Nope, they can’t. This host is down. You can’t access its printer.
Another problem is the fact that when you have a peer-to-peer network, there’s no real centralised place to save files. We could have storage on this system. We could have maybe some other shared storage on this system. Maybe some other shared storage on this system. Maybe even some over here on this system. When it comes time to back up and protect critical company data, you’ve got to go to each one of these hosts and back up their data. You may not know exactly where folks are saving it. It could be all over the place on the hard drives.
We have another classification within the host roles category, and that is a client/server network. In a client/server network, unlike a peer-to-peer network, network hosts have specific roles assigned to them. In a client/server network, you have certain systems, certain hosts that are assigned to be servers. What does a server do? A server provides network resources. You also on a client/server network have clients. A client does not provide network services. It consumes network services. What does that mean? That means in a client/server network these client workstations will not have shared storage. They will not have shared printer. All those services are provided down here by the server. It might have a humongous hard drive. Maybe we have a printer that this server manages. Et cetera. Server provides the resources. The clients simply use the resources.
To do that, we have different operating systems implemented. Recall, with a peer-to-peer network, everybody had the same operating system. Not so in a client/server network. Client workstations have generic operating systems that provide functionality such as being able to run applications, to do word processing, spreadsheets, et cetera, and software that allows them to connect to the server. Examples would be Windows XP, maybe Windows 2000 Professional. Those are client workstation operating systems. The server, on the other hand, has special optimised operating systems, and these operating systems aren’t designed for client type tasks such as word processing and spreadsheets. Although some of them will do it, that’s not their main job. For example, server operating systems include NetWare, Linux, Windows 2000, and Server 2003. These operating systems are designed to provide these network resources.
The benefit to this type of network is first of all that it’s very highly scalable. What does that mean? That means it’s very easy to expand the size of the network. It’s very easy to add more clients. It’s very easy to add more servers. Client/server networks are also much easier to support. That’s because services are centralised. If folks are having a problem accessing their files, you know where to look. It’s on the server. You need to check out the storage on the server. If folks are having trouble printing, you know where to look, because the service is provided by the server. Backup is also a lot easier. Instead of having to back up individual workstations, they’re saving their data over here in the storage system on the server itself.
You’re still using your relatively inexpensive client operating systems up here just like with a peer-to-peer network, such as Windows XP, Windows 2000 Professional. The expense comes down here. These operating systems for the most part are relatively expensive. The exception, of course, is Linux. NetWare is relatively expensive. Windows 2000 Server is very expensive. Server 2003 is relatively expensive as well.
The other thing is that this type of network takes a lot of planning beforehand. If it were a peer-to-peer network, you just slap things together. You set up your systems, and away you go. You don’t do that with a client/server network. You pre-plan everything. You decide which servers are going to host which services and you decide where they’re going to be placed on the network, et cetera.
That’s how we categorise computer networks by host role. Now let’s look at a different way of categorising computer networks, and that is geography. The first category is that of a local area network. A local area network resides within a small geographic area. An example of a local area network would be the network inside a particular company’s office. It has multiple floors, but these are all connected by a network medium in some way. This comprises a local area network. It could be multiple buildings, in fact. You could have a second building over here, several floors, computer systems, and these are all interconnected in some fashion. It’s still a local area network because the geography separating the hosts is relatively small. You could even have other buildings over here with their own networks. For instance, perhaps a college campus. There’s a building here and a building here, and then you connect all the networks together. When you do that, by the way, that’s called an inter-network. However, this is still a LAN because they’re geographically close together.
It’s also possible to have a computer network where the networks and the hosts are very widely distributed geographically. When we have that situation, we’re talking about a wide area network. Basically, a wide area network is a group of interconnected LANs, local area networks, that are separated geographically. For example, suppose we have a company named XYZ Corp. XYZ has an office over here in New York City. They have a local area network. This company also has an office down here in Houston, Texas. Then they also have an office up here in Portland, Oregon. Users in these different locations occasionally need to have access to information at these various different sites. To make that possible, we’ve connected them in some way so that these local area networks are inter-networked together, making a very large inter-network, or a wide area network.
When you set up a WAN, basically the way it’s done is that you make sure that the resources that the users need are located on the local LAN. For example, if someone needs to open a file here in New York City, you don’t want them to have to go clear down here, transverse the wide area network to Houston to open up a file and bring it back here and work on it over here in New York City. Then, when you save you have to go clear across the country again just to save on a server located down here. Instead, you’ll locate the information resources that each local area network needs here locally, so you’d set up a server here in New York City with the information that the users here in New York City would need.
Likewise, you wouldn’t want to store your user accounts over here in Portland so that someone who needs to log in to the network down here in Houston has to go across the WAN link clear over here to Portland just to authenticate. The benefit of the WAN is that, if the situation arises, say a user here needs access to a document that’s being worked on by a product team over here in Portland, they can use the wide area network to get that information and open it over here.
Those are the two different classification or categorising networks by their size and their geographic proximity. There’s one more categorization I want to talk about here, and that is the type of signalling used by the network. There’s basically two types. There’s baseband signalling and then there’s broadband signalling. This right here represents our network medium, whether it’s a wire or whether it’s a piece of fibre optic cable or what have you. With baseband signalling, you can have one signal at a time on the network medium, and that signal uses the entire network medium all at the same time. Broadband signalling, on the other hand, divides the network medium into multiple channels. With broadband signalling, you can have multiple signals all being transmitted on the same network medium at exactly the same time.
You want a good example of broadband signalling, think of your cable TV system. If you have cable TV, you have channel two, four, five, six, seven, et cetera. You have multiple signals all using the same cable at the same time. Baseband signalling, in the old days I used to use the example of a telephone. Now telephone systems are starting to use broadband signalling to allow you to have DSL and your telephone conversations at the same time, but in the old days telephones used to use baseband signalling. In other words, you picked up the phone receiver and you made a telephone call. While you were using the phone, someone else on a different extension in the same house couldn’t lift up the phone and make a phone call at the same time because you could only have one signal on the wire at the time.
Those are the two different types of signalling that’s used-