Peer-to-Peer vs. Client/Server Architecture

Peer-to-Peer vs. Client/Server Architecture
As you learned earlier in this chapter, the purpose of networking is to share resources. How this
is accomplished depends on the architecture of the network operating system software. The two
most common network types are peer-to-peer and client/server.

If you were to look at an illustration of a group of computers in a LAN, it would be impossible to determine if the network was a peer-to-peer or a client/server environment. Even a videotape of this same LAN during a typical workday would reveal few clues as to whether it is peer-to-peer or client/server. Yet, the differences are huge. Since you can’t see the differences, you might guess correctly that they are not physical but logical.

Physical vs. Logical Concepts
Throughout this book, you’ll see us refer to physical and logical networking topics. Generally speaking, when we’re referring to the physical aspects of a network, we’re referring to some aspect of the network that you can touch or that has physical substance (like electrons, electrical pulses, or the way cables are run). That is, they exist in the physical world. Logical concepts, on the other hand, are more imaginary and esoteric and deal with things like how data flows in a network. So, when we’re describing something as either physical or logical in nature, you’ll understand how those terms apply.

Peer-to-Peer Network
In peer-to-peer networks, the connected computers have no centralized authority. From an authority viewpoint, all of these computers are equal. In other words, they are peers. If a user
of one computer wants access to a resource on another computer, the security check for access rights is the responsibility of the computer holding the resource. Each computer in a peer-to-peer network can be both a client that requests resources and a server
that provides resources. This is a great arrangement, provided the following conditions are met: 
 

• Each user is responsible for local backup.
•   Security considerations are minimal.
• A limited number of computers are involved. 

Networks that run Windows 95/98 as their network operating system and networks using Windows NT, 2000, or XP in a workgroup are considered peer-to-peer networks. Figure 1.3 shows an example of a peer-to-peer network. Peer-to-peer networks present some challenges. For example, backing up company data becomes an iffy proposition. Also, it can be difficult to remember where you stored a file. Finally, because security is not centralized, users and passwords must be maintained separately on each machine, as you can see in Figure 1.3. Passwords may be different for the same users on different machines 

(or for different resources on Windows 9x machines).

Client/Server Network
In contrast to a peer-to-peer network, a client/server network uses a network operating system designed to manage the entire network from a centralized point, which is the server. Clients make requests of the server, and the server responds with the information or  ccess to a resource.

Example:

A peer-to-peer network

Client/server networks have some definite advantages over peer-to-peer networks. For one thing, the network is much more organized. It is easier to find files and resources because they are stored on the server. Also, client/server networks generally have much tighter security. All usernames and passwords are stored in the same database (on the server), and individual users can’t use the server as a workstation. Finally, client/server networks have better performance and can scale almost infinitely. It is not uncommon to see client/server networks with tens of thousands of workstations. Figure 1.4 shows a sample client/server network. Note that the server now holds the database of user accounts, passwords, and access rights.
Note that today’s networks are very often hybrids of the peer-to-peer model and the client/server model. Clients of early Novell NetWare networks, for example, had no ability to share  their resources, not that they had many worth sharing, for the most part. Conversely, today’s Microsoft and Apple networks, for example, have well-defined servers. They also allow the simultaneous sharing of resources from lesser devices that run what are considered workstation operating systems, which are capable of fewer inbound connections but are running the server service nonetheless. Purists shun the less organized mixture of this resource sharing among servers and clients alike, but the reality is that most networks would be worse off for losing this capability.

Example:

A client/server network