next up previous
Next: Internetworking Up: Network Topologies and Access Previous: Network Topologies and Access

The Ethernet

We now look very briefly at the Ethernet, a local-area packet-switched network technology invented at Xerox PARC in the early 1970s that originally used the bus topology with the CSMA/CD protocol. The channel used for communication was a coaxial cable called the ether, whose bandwidth is 10 Mbps and maximum total length 1500 meters.

Each connection to the ether has two major components. A transceiver connects directly to the ether cable, sensing and sending signals to the ether. A host interface connects to the transceiver and communicates (as a device) with the computer (usually through the computer bus).

Each host interface attached to a computer is assigned a 48 bit Ethernet address also called a physical address. This address is used to direct communication to a particular machine. Vendors purchase blocks of physical addresses and assign them in sequence as they manufacture Ethernet interface hardware.

Each packet transmitted along an ether is called an Ethernet frame. A frame is a maximum of 1536 bytes and contains:

a preamble (64 bits or 8 octets), used for synchronization,

source and destination addresses (6 octets),

packet type (2 octets), used by higher layers,

data (46 to 1522 octets), and

Cyclic Redundancy Code (CRC) (4 octets), which is a function of the data in the frame and is computed by both the sender and the receiver.

The first three fields of a frame form its header.

A destination address in the packet may be the physical address of a single machine, a multicast address of a group of nodes in the network, or the network broadcast address (usually all 1's). A host interface picks up a frame if the destination address is:

the physical address of the interface

a multicast address of the group to which the host belongs, or

one of the alternate addresses specified by the operating system.

the broadcast address

Today, Ethernet consists of twisted pairs connecting to a central hub. The twisted pairs can come in two configurations: (a) a single physical link for carrying traffic in both directions, or (b) separate physical links for incoming and outgoing traffic. The hub behaves as a switch, directing an incoming message to its destination(s).

next up previous
Next: Internetworking Up: Network Topologies and Access Previous: Network Topologies and Access
Prasun Dewan 2006-02-02