Network Design

Computer networks also differ in their design. The two types of high-level network design are called client-server and peer-to-peer. Client-server networks feature centralized server computers that store email, Web pages, files and or applications. On a peer-to-peer network, conversely, all computers tend to support the same functions. Client-server networks are much more common in business and peer-to-peer networks much more common in homes.

In computer networking, topology refers to the layout of connected devices. This article introduces the standard topologies of networking.

Topology in Network Design

Think of a topology as a network's virtual shape or structure. This shape does not necessarily correspond to the actual physical layout of the devices on the network. For example, the computers on a home LAN may be arranged in a circle in a family room, but it would be highly unlikely to find a ring topology there.

Network topologies are categorized into the following basic types:

• bus
• ring
• star
• tree
• mesh

Bus Topology

Bus networks (not to be confused with the system bus of a computer) use a common backbone to connect all devices. A single cable, the backbone functions as a shared communication medium that devices attach or tap into with an interface connector. A device wanting to communicate with another device on the network sends a broadcast message onto the wire that all other devices see, but only the intended recipient actually accepts and processes the message.

Ethernet bus topologies are relatively easy to install and don't require much cabling compared to the alternatives. 10Base-2 ("ThinNet") and 10Base-5 ("ThickNet") both were popular Ethernet cabling options many years ago for bus topologies. However, bus networks work best with a limited number of devices. If more than a few dozen computers are added to a network bus, performance problems will likely result. In addition, if the backbone cable fails, the entire network effectively becomes unusable.

Ring Topology

In a ring network, every device has exactly two neighbors for communication purposes. All messages travel through a ring in the same direction (either "clockwise" or "counterclockwise"). A failure in any cable or device breaks the loop and can take down the entire network.

To implement a ring network, one typically uses FDDI, SONET, or Token Ring technology. Ring topologies are found in some office buildings or school campuses.

Star Topology

Many home networks use the star topology. A star network features a central connection point called a "hub" that may be a hub, switch or router. Devices typically connect to the hub with Unshielded Twisted Pair (UTP) Ethernet.

Compared to the bus topology, a star network generally requires more cable, but a failure in any star network cable will only take down one computer's network access and not the entire LAN. (If the hub fails, however, the entire network also fails.)

Tree Topology

Tree topologies integrate multiple star topologies together onto a bus. In its simplest form, only hub devices connect directly to the tree bus, and each hub functions as the "root" of a tree of devices. This bus/star hybrid approach supports future expandability of the network much better than a bus (limited in the number of devices due to the broadcast traffic it generates) or a star (limited by the number of hub connection points) alone.

Mesh Topology

Mesh topologies involve the concept of routes. Unlike each of the previous topologies, messages sent on a mesh network can take any of several possible paths from source to destination. (Recall that even in a ring, although two cable paths exist, messages can only travel in one direction.) Some WANs, most notably the Internet, employ mesh routing.

A mesh network in which every device connects to every other is called a full mesh. As shown in the illustration below, partial mesh networks also exist in which some devices connect only indirectly to others.

Summary

Topologies remain an important part of network design theory. You can probably build a home or small business computer network without understanding the difference between a bus design and a star design, but becoming familiar with the standard topologies gives you a better understanding of important networking concepts like hubs, broadcasts, and routes.

The Open Systems Interconnection (OSI) reference model has been an essential element of computer network design since its ratification in 1984. The OSI is an abstract model of how network protocols and equipment should communicate and work together (interoperate).

The OSI model is a technology standard maintained by the International Standards Organization (ISO). Although today's technologies do not fully conform to the standard, it remains a useful introduction to the study of network architecture.

The OSI Model Stack

The OSI model divides the complex task of computer-to-computer communications, traditionally called internetworking, into a series of stages known as layers. Layers in the OSI model are ordered from lowest level to highest. Together, these layers comprise the OSI stack. The stack contains seven layers in two groups:

Upper layers -

7. application
6. presentation
5. session

Lower layers -

4. transport
3. network
2. data link
1. physical

Upper Layers of the OSI Model

OSI designates the application, presentation, and session stages of the stack as the upper layers. Generally speaking, software in these layers performs application-specific functions like data formatting, encryption, and connection management.

Examples of upper layer technologies in the OSI model are HTTP, SSL and NFS.

Lower Layers of the OSI Model

The remaining lower layers of the OSI model provide more primitive network-specific functions like routing, addressing, and flow control. Examples of lower layer technologies in the OSI model are TCP, IP, and Ethernet.

Benefits of the OSI Model

By separating the network communications into logical smaller pieces, the OSI model simplifies how network protocols are designed. The OSI model was designed to ensure different types of equipment (such as network adapters, hubs, and routers) would all be compatible even if built by different manufacturers. A product from one network equipment vendor that implements OSI Layer 2 functionality, for example, will be much more likely to interoperate with another vendor's OSI Layer 3 product because both vendors are following the same model.

The OSI model also makes network designs more extensible as new protocols and other network services are generally easier to add to a layered architecture than to a monolithic one.