The need for bridging devices arises with the need to communicate with computers located beyond a particular local area network (LAN) . Although all the clients (computers) attached to a LAN need not be located in the same room, there are limitations on the distance between clients as well as on the number of clients that can be attached to a single network segment. Bridging devices are used to overcome these limitations and facilitate communication among machines on different floors, different buildings, different cities, and different countries.
Bridging devices are available in a variety of configurations to interconnect multiple local area network segments. The choice of bridging device depends on distance, traffic volume, and complexity of the communication pathways between sites. Commonly used bridging devices include repeaters, bridges, routers, and gateways.
The simplest bridging device is known as a repeater. As messages travel over increasing distances, their signals become weak and distorted. A repeating device extends the distance over which clear communication can take place by regenerating messages. For example, repeaters can be used to facilitate communication among computers in a LAN that spans several floors of a building.
The LAN cable for each floor is connected to a repeater, which is sometimes called a hub. Additional cabling connects a repeater to other repeaters on adjacent floors. As messages travel from floor to floor, their signal strength is maintained because the repeaters regenerate them.
Repeaters are limited as bridging devices because they operate at the physical network layer and simply pass on all the bits that they receive. Repeaters do not distinguish between messages intended for clients on the same floor and those intended for clients on different floors. Repeaters also retransmit messages with errors and signals resulting from collisions when clients attempt to send simultaneous messages. If traffic volume is heavy, performance on the network will deteriorate.
A device known as a bridge is used to reduce network traffic by filtering messages. Bridges, unlike repeaters, do not retransmit every signal that they receive. Bridges operate at the data link layer of the networking hierarchy. They filter messages using the hardware or MAC (medium access control) addresses of the PCs attached to the local network. A bridge retransmits only correct messages with destination addresses that fall outside the network segment from which the message originated.
A bridge, for example, can be used to connect LAN segments located in two different buildings. Although the two network segments function as a single LAN, the bridge limits traffic between buildings to messages involving client PCs actually located in different buildings. Performance for the entire LAN will be better because a pair of clients in each building will be able to exchange messages without interfering with communications in the other building. Network designers use bridges to improve performance on busy LANs by dividing the network into segments connected by bridges and assigning computers that frequently exchange messages to the same segment.
To filter messages, bridges must know the location of the client computers attached to their network segments. When a bridge receives a message, it recovers the hardware address of the sending computer and adds it to a table that associates computers with network segments. After each PC on the network has sent one message, the bridge will have a complete record of PCs and their locations.
This complete table of computers and location addresses is crucial to the operation of the bridge. Bridges compare the destination addresses in messages to their tables to determine when to retransmit a message. If the destination address belongs to the segment over which the message arrived, the destination computer has already received the message and the bridge does not need to forward the message. If the destination address does not belong to the segment over which the message arrived, the bridge will forward a copy of the message to the other network segment.
The great advantage of a bridge is that it is a plug-and-play device that requires no set-up by a network administrator. Bridges are very effective for small networks involving just a few segments. As the number of segments in a network increases, redundant multiple paths between distant LAN segments become an important issue. Redundancy in a network means alternate pathways between network locations are available in case of a failure or congestion in one part of a network. Unfortunately, the rules by which bridges operate restrict the effective pathways between network segments. Consequently, with an increase in both the number of segments in an organization's network and the need to make connections across networks (e.g., the Internet), bridges become less effective than routers as the bridging device of choice.
A router differs from a bridge because it operates at the network layer of the network model. The addresses used by routers are network addresses such as the familiar "dotted-decimal" addresses found on the Internet. Each message packet must specify a network destination address (e.g., 220.127.116.11). In a complex network with multiple alternate paths between locations, the primary task of the routers on a network is to use the destination address of a message to determine the best path between locations for a message to follow.
Packets traveling through a network are passed from router to router until they reach a router that is attached to the same local network as the destination. When a packet arrives at a router, the router first determines if it is addressed to a device on its local network. If not, the router chooses the next router to receive the packet in its progress across the network.
The data that a router will use to make its forwarding decisions are contained in a routing table. The routing table contains a list of destination addresses and the "next-hop" router appropriate for each of those destinations. The minimum set of entries in a routing table includes an address to identify devices on the router's local network and a default address for forwarding all other packets. More complex tables present alternative pathways so that a router can choose to forward a message along the most direct, the least congested, or the fastest pathway. Routers in a network are able to exchange data to update their routing tables with information about failures, congestion, or new paths in the network.
All the routers on the Internet operate according to the Internet Protocol (IP) . A different kind of bridging device, called a gateway, is needed to interconnect two networks that use different network layer protocols. When an IP router communicates with an adjacent IP router, it is only necessary to package the data for the shared network protocol in a data-link layer frame. However, when an IP router needs to communicate with, for example, an IBM SNA router, it will also be necessary to replace the IP network data with appropriate SNA network data.
The task of translating network layer descriptions for different networks is performed by a gateway. The gateway receives messages from one network, removes the network layer information, adds network layer information formatted for the second network, packages the message in a data-link layer frame and forwards it to the second network.
see also E-commerce; Internet; Networks; Telecommunications; World Wide Web.
Comer, Douglas. Computer Networks and Internets, 2nd ed. Upper Saddle River, NJ: Prentice Hall, 1999.
FitzGerald, Jerry, and A. Dennis. "High-Speed LAN and Backbone Networks." In Business Data Communications and Networking, 6th ed. New York: Wiley, 1999.
Kurose, James, and K. Ross. Computer Networking: A Top-Down Approach Featuring the Internet. Reading, MA: Addison-Wesley, 1999.