Network Range & Topology

Networks can be categorized based on several key characteristics. These characteristics can be combined to form networks of varying sizes as well as layout. A network's range classifies the geographical area for which the network spans and reaches. There are two general distinctions when it comes to describing networks based on their range.

LAN

A LAN or Local Area Network is a geographically small network consisting of nodes, computers, or other networkable devices. All of the devices in a LAN are usually found within the same building or structure. LAN's are used to segregate and organize large networks into smaller, more manageable pieces. These networks tend to be owned and operated by the same people who own the nodes that connect to the network.

WAN

A WAN or Wide Area Network is a geographically larger network than a LAN. A WAN is usually owned/maintained by a larger company. These companies may use the WAN for their own corporate use or they may lease out the use of their network for communication. Telecommunication companies such as Time Warner and BellSouth run WANs that provide a method for people and other companies to join to other LAN's and connect to global communications such as the Internet.

 

In addition to the geographical range of a network, the physical layout of the network is also used to classify networks. The physical layout represents the way in which each node connects together via some connection medium (see Network Mediums).

Star Network

A star network is a network by which computers are connected by one central device known as a concentrator. The concentrator manages and controls all functions of the network. In addition, the concentrator acts as a repeater for the network. A repeater is a device that strengthens the signal of the communication as it travels over the appropriate medium.

Fig. 3: Star network

It is easy to install and connect all the nodes of a star network. If it is necessary to work on or disconnect a particular node on the network, there is no interruption of service to the rest of the network. Also, the concentrator makes finding a troublesome or malfunctioning node on the network easy. The faulty node can be traced back via the port on the concentrator and the connection medium. While the loss of one node on the network will not affect the rest of the nodes on the network, if the concentrator fails, the entire network will be disabled. Star networks require more cable than a bus/linear topology and the addition of the concentrator makes a star network more expensive than a bus network.

Ethernet switches and hubs are both a good example of a concentrator. Star networks are rarely used by themselves, rather they are put together to form a Tree Network.

Tree Network

A tree topology can be thought of as a collection of star networks arranged in a heirarchy. Like a traditional star topology, each node in a tree network is connected to its own concentrator. However, concentrators from each star network are connected in a heirarchal manner.

Fig. 4: Tree network

Like a star network, a point of failure on a transmission line can separate a node from the network and the rest of the network will continue to work. If a concentrator gets disconnected, all connected nodes are separated from the more massive tree network. The nodes on the separated star network, however, can still continue to communicate with one another as long as the concentrator is still functioning.

Bus Network

One of the earliest networks designed was a bus network. In a bus network, a set of clients are connected via a single shared connection line known as a bus. This network is very similar to the system bus in your computer. While bus networks are the simplest way to connect devices together, problems occur when two devices want to communicate at once. Bus networks typically use a collision detection/prevention method known as CSMA (Carrier Sense Multiple Access) to make sure all data is communicated properly.

Fig. 1: Bus network

A bus network topology has several advantages. It is one of the easiest and cheapest topologies to implement and extend. To add a new device, one must simply tap into the communication line and he or she will be ready to go. The only equipment needed is an interface card for each node and one communication line to establish the network. The small amount of necessary equipment keeps the cost low. In addition to the small amount of required equipment, a bus network is also well suited for quick, temporary network setups. Also, if one node fails along the bus, it does not bring down the entire network.

While there are some great advantages to a bus topology, there are also some disadvantages. Since there is only one line of communication in use, a line break will cause the network to go down completely. In addition, that line can only run a limited distance and support a limited amount of nodes before one can see the network begin to slow to a crawl. It can be difficult to troubleshoot problems on the network because there is no way to isolate an individual node or group of nodes on the network. Due to the difficulty in troubleshooting and the relatively small number of nodes that can be supported, maintenance cost may be higher in the long run. The increased troubleshooting time as well as the creation of separate networks to handle more nodes can increase cost.

Time Warner Cable's cable broadband service is a good example of a bus network

Ring Network

In a ring network, nodes are connected in a ring configuration. Each node has two neighbors. Data can travel either clockwise or counterclockwise through the ring, but all data travels in only one direction. For information to be sent from one node to another, it must be sent around the ring and through each node until it reaches its appropriate destination.

Fig. 2: Ring network

Since all nodes have equal access to the network, there is a minimal impact on the performance of the network as devices are added. No one node can dominate communication on the network. Unfortunately, due to the design, if one device fails on the network, there is a good chance that other nodes will also be effected and fail. In addition, the entire network may stop working all together. Also there is additional cost incurred due to the type of interface cards and other controlling devices present in a typical ring setup. In addition, communication between the devices can be complicated depending on the implementation of the ring network.