Peer-to-Peer Versus Client/Server Networks

Networks use two basic models: peer-to-peer and client/server. The model used by an organization depends on the role of the network and what the users require from it. You will probably encounter both network models; therefore, you need to understand how these models work as well as their strengths and weaknesses.

The Peer-to-Peer Networking Model

Peer-to-peer networking is a low-cost, easily implemented network solution that is generally used in small network environments that need to share a few files and maybe some hardware, such as printers. As its name suggests, on a peer-to-peer network all systems are equal, or peers. Each system can share hardware or files and access the same things on other systems.

A peer-to-peer network offers no centralized data storage or centralized control over the sharing of files or resources. In a sense, everyone on a peer-to-peer network is a network administrator and can share resources as they see fit. They have the option to grant all users on the network complete access to their computers, including printers and files, or they can choose not to share anything. Figure 1.1 shows an example of a peer-to-peer network.

The peer-to-peer model works quite well on networks that have 10 or fewer computers, but as a network grows, it becomes more complicated. Peer-to-peer networking is often referred to as decentralized networking because the network files, data, and administration are not handled from a central location. This arrangement can lead to huge problems, especially in large networks. For example, locating specific files can become difficult because the files might be on multiple computers. Data backup cannot be performed from a central location; each computer must be backed up individually. Decentralized computing can also be difficult in terms of network security because security is controlled by individual computer users rather than being administered from a central location. This decentralized security model requires that each user have a user ID and password defined on each and every system that he or she will access. With no way of synchronizing passwords between the systems, this can quickly become a problem. Many users have problems remembering just one or two passwords—let alone a dozen.

Given the complexity and drawbacks of using peer-to-peer networking, you might wonder why anyone would use it. Many small companies begin with a peer-to-peer network because it's the easiest and least expensive type to install. After the networks grow too big, they switch to the client/server model, which is discussed in the next section.

Advantages of Peer-to-Peer Networks

The following are some of the advantages of using the peer-to-peer networking model:

• Cost— Because peer-to-peer networking does not require a dedicated server, such networks are very cost-effective. This makes them an attractive option in environments where money is tight.

• Ease of installation— The built-in support for peer-to-peer networking in modern operating systems makes installing and configuring a peer-to-peer network a straightforward process.

Disadvantages of Peer-to-Peer Networks

The following are some of the disadvantages of using the peer-to-peer networking model:

• Security— In a decentralized model, a networkwide security policy cannot be enforced from a server; rather, security needs to be applied to each computer and resource individually.

• Data backup— Because files and data are located on individual computers, each system must have its data backed up individually.

• Resource access— In a decentralized approach, it can be difficult to locate resources on the network. Printers and files may be distributed among numerous computer systems.

• Limited numbers of computers— Peer-to-peer networking is effective only on small networks (fewer than 10 computers).

As you can see, the disadvantages of peer-to-peer networking outweigh the advantages. Therefore, client/server networks are far more popular in corporate or business environments than peer-to-peer networks.

The Client/Server Networking Model

Client/server networking—or server-based networking, as it is commonly called—is the network model you are most likely to see in the corporate world. The server-based network model is completely scalable, allowing additional computers or other networked devices to be added with little difficulty. Perhaps the greatest benefit of this model is that it allows for centralized management of all network services, security, and streamlined backup procedures. Figure 1.2 shows an example of a client/server network.

As you may have gathered, two different types of computers are required for the server-based model: the client and the server. Figure 1.3 shows the relationship between client and server computers. These two computer systems are often very different from each other, and each plays a unique role on a network.

Servers

Servers are the workhorses of the network. They spend their time responding to the numerous requests that come from client computers, such as requests for files, network authentication, and access to shared hardware resources. Network administration—including network security, backups, and network monitoring—is done from the server.

To be able perform their functions, server computers require additional resources and computing power. Server systems often use specialized hardware and software in fault-tolerant configurations to ensure that they remain operational. When a server fails and goes offline, it is unable to respond to requests from client systems and its functions are unavailable. This situation can be frustrating for users and very costly for an organization.

In addition to requiring specific hardware, servers also require a network operating system. A network operating system stands above ordinary desktop operating systems; it has unique features and functions that allow an administrator to manage, monitor, and administer the data and resources of the server as well as the users who connect to it. In addition, network operating systems are designed to be resilient in case of the kind of downtime discussed earlier. The most common network operating systems used today are the Microsoft Windows network operating systems (including Windows NT, Windows 2000, and Windows XP/Windows .Net), Unix, Linux, and Novell NetWare. Knowledge of these operating systems is an important element of the Network+ exam, so detailed coverage of them is provided in Chapter 9, “Network Operating Systems and Clients.”

A network may have a single server that offers more than one network service or hundreds of servers, each performing a dedicated task. For instance, one server might be used only to authenticate users and another might be used to store an applications database. Some of the most common roles for dedicated servers include acting as file and print servers, application servers, Web servers, database servers, firewall servers, and proxy servers.

Client Computers

Client computers are the other half of the client/server model. Client computers connect to the network and access the resources of the server. Software is needed to allow the client to connect to the network, although the need for networking has become so fundamental that the client software functionality is now built in to desktop operating systems. More coverage of the role and requirements of client computers is provided in Chapter 9.

Advantages of Client/Server Networking

The following are some of the advantages of the client/server networking model:

• Centralized management and security— The ability to manage the network from a single location is the biggest advantage of the client/server model. From a server, you can perform backups of all data, share resources and control access to those resources, manage user accounts, and monitor network activity.

• Scalability— In a server-based network administrators can easily add computers and devices. In addition, the network is not restricted to a small number of computers. In a client/server network, the number of clients is limited by factors such as licensing and network capacity rather than by the operating system's ability to support them.

• Simplified backups— On server-based networks, files and folders typically reside in a single location or a small number of locations and are therefore easier to back up than the files on a peer-to-peer network. Scheduling backups to occur at regular intervals is simple.

Disadvantages of Client/Server Networking

The following are some of the disadvantages of the client/server networking model:

• High cost— A server-based network requires additional hardware and software, so it can be a costly venture. The costs of the client/server model include the costs of the network operating system and at least one server system, replete with specialized server hardware. Also, because the client/server model can support far more systems than the peer-to-peer model, networking devices such as hubs, routers, and switches are often needed.

• Administration requirements— Client/server networks require additional administrative skills over those required on a peer-to-peer network. In particular, the technical capabilities of the administrator need to be greater. Organizations that use the server-based model often need technically skilled people to manage and maintain the network and the servers.

• Single points of failure— In a client/server model, the client systems depend on servers to provide network services. If the server fails, the clients can't access the services that reside on the server. Great effort and expense are needed to ensure the high availability of network servers.

Given the limitations of the peer-to-peer network design, such networks are used in only a few situations. On the other hand, the client/server networking model is versatile, and its shortcomings are overshadowed by its capabilities and advantages. You will spend most of your time working with server-based networks of all shapes and sizes.

Distributed and Centralized Computing

Although they're less of an issue than in the past, you need to be familiar with two important networking concepts: distributed and centralized computing. These concepts are not directly related to the server-based/peer-to-peer discussion, although by definition a peer-to-peer system is a distributed computing model.

The terms distributed and centralized computing describe the location on a network where the processing takes place. In an environment such as a mainframe, the processing is performed on a centralized system that also stores all the data. In such a model, no data processing or data storage occurs on the client computer. In contrast, in a distributed processing environment, processing is performed in more than one place. If a network has servers and workstations, processing can take place on the server or on the client.

It is relatively unusual for a company to have just a centralized computing environment. A company is far more likely to have a server-based network, which would fall under the banner of distributed computing, and perhaps a mainframe that is accessed from the same PCs as the server-based network, which would fall under the banner of centralized computing. A good example of such an environment might be a company that books hotel reservations for customers, in which the booking system is held on a mainframe but the email system used to correspond with clients is held on a PC-based server and accessed through client software on the PCs.

Regardless of whether a network uses a centralized or distributed model, networks can be placed into categories that define the geographic location over which they are spread. The three accepted terms for these categories are LAN, WAN, and MAN.