Summary of Module 8

Summary
This page summarizes the topics discussed in this module.


Ethernet is a shared media, baseband technology, which means only one node can transmit data at a time. Increasing the number of nodes on a single segment increases demand on the available bandwidth. This in turn increases the probability of collisions. A solution to the problem is to break a large network segment into parts and separate it into isolated collision domains. Bridges and switches are used to segment the network into multiple collision domains.

A bridge builds a bridge table from the source addresses of packets it processes. An address is associated with the port the frame came in on. Eventually the bridge table contains enough address information to allow the bridge to forward a frame out a particular port based on the destination address. This is how the bridge controls traffic between two collision domains.

Switches learn in much the same way as bridges but provide a virtual connection directly between the source and destination nodes, rather than the source collision domain and destination collision domain. Each port creates its own collision domain. A switch dynamically builds and maintains a Content-Addressable Memory (CAM) table, holding all of the necessary MAC information for each port. CAM is memory that essentially works backwards compared to conventional memory. Entering data into the memory will return the associated address.

Two devices connected through switch ports become the only two nodes in a small collision domain. These small physical segments are called microsegments. Microsegments connected using twisted pair cabling are capable of full-duplex communications. In full duplex mode, when separate wires are used for transmitting and receiving between two hosts, there is no contention for the media. Thus, a collision domain no longer exists.

There is a propagation delay for the signals traveling along transmission medium. Additionally, as signals are processed by network devices further delay, or latency, is introduced.

How a frame is switched affects latency and reliability. A switch can start to transfer the frame as soon as the destination MAC address is received. Switching at this point is called cut-through switching and results in the lowest latency through the switch. However, cut-through switching provides no error checking. At the other extreme, the switch can receive the entire frame before sending it out the destination port. This is called store-and-forward switching. Fragment-free switching reads and checks the first sixty-four bytes of the frame before forwarding it to the destination port.

Switched networks are often designed with redundant paths to provide for reliability and fault tolerance. Switches use the Spanning-Tree Protocol (STP) to identify and shut down redundant paths through the network. The result is a logical hierarchical path through the network with no loops.

Using Layer 2 devices to break up a LAN into multiple collision domains increases available bandwidth for every host. But Layer 2 devices forward broadcasts, such as ARP requests. A Layer 3 device is required to control broadcasts and define broadcast domains.

Data flow through a routed IP network, involves data moving across traffic management devices at Layers 1, 2, and 3 of the OSI model. Layer 1 is used for transmission across the physical media, Layer 2 for collision domain management, and Layer 3 for broadcast domain management.