In an Ethernet network, each computer is called a node, which has a unique identifying address. We identify our computers by giving them a specific name, such as Spongebob, SERVER1 or FINANCE1. However, on an Ethernet network, computers communicate using strings of ones and zeroes which is known as binary code. Ethernet uses special 48-bit addresses known as MAC addresses, which is used to identify individual computers on a network. Each network card in a computer contains a unique MAC address.

There are a possible 281, 474, 976, 710, 656 MAC addresses. An Ethernet network could consist of more than 281 trillion machines. However, there are factors that limit networks to a much smaller size. To make MAC addresses easier for network technicians, addresses are written in a hexidecimal notation.

MAC addresses are to Ethernet networks as I/O addresses are to individual computers. I/O addresses provide a way for the CPU to identify each component in a computer by giving every component a unique I/O address. If two components are present and share the same I/O address, neither will function properly. Commands sent by the CPU to one of the components causes a conflict because both components attempt to respond to the command, which results in them talking over each other and the CPU receiving unintelligible reponses. This is a good example of what happens if two componenets on a network share the same MAC addresses.

The IEEE has taken great care in ensuring that no two network interface cards (NICs) share the same MAC address. Network interface card manufacturers must apply to the IEEE for a block of MAC addresses. To determine the MAC address of a specific NIC, the diagnostic utility that came with the NIC card could be used. On Windows 95/98 running TCP/IP protocol, you can run WINIPCFG. And, on Windows NT, you can run IPCONFIG. But if you like, you can look at your NIC. Many manufacturers place a MAC address label on the cards.