Ethernet Addresses refers to the unique physical/hardware address assigned or hardcoded on to the Ethernet interface card in 48 bits machine readable form. Physical ethernet addresses are meant to be different for each interface card and moving the hardware to a new machine or replacing the existing card with a new one will change the physical address of the host computer.
The ethernet interface card could recognize a set of addresses which are of interest to it. For e.g., it has to recognize its physical address, the network broadcast address and the multicast addresses all sent to it. Usually, when the machine is booted, the NOS (network operating system) will supply these set of addresses to the interface card to recognize. The interface now scans each transmission, passing on to the host only transmissions designated for one of the specified addresses.
Ethernet frame format:
Ethernet frame format is as shown above. Ethernet transmits the data signals as frames of data. Ethernet frames are not smaller than 64 octets or larger than 1518 octets. The preamble bits (64 bits) consists of alternate 0s and 1s to help the receiving ands to synchronize the frames. The destination and source addresses are each 48 bits long. The frame data carries the message. The CRC stores the cyclic redundancy check sum values computed by the sender as a function of the frame data. The receiver re-computes the CRC value and checks for errors. The frame type field indicates which protocol is being used by the frames and this which corresponding protocol is to be used by the host computer from a set of local protocol suites. The type field in ethernet frames makes ethernet self identifying since they enables the host to apply the appropriate protocol and also allow the ethernet medium to carry a collection of frames employing different protocols that are directed to be used by the appropriate protocol set in host computers.