Join Timothy Pintello for an in-depth discussion in this video Exploring physical addressing, part of Foundations of Networking: IP Addressing.
- Now that we've looked at some of the numbering systems used with network addressing, let's go ahead and look at some actual types of network addressing. Network addressing comes in two types, physical addressing and the logical addressing. Let's look at the physical addressing first. The physical address refers to the actual physical address assigned to each device on a network. This physical address is set at the factory when the device being used is actually manufactured. The physical address is not something that you can go in and add manually yourself after the fact.
The physical address, you'll oftentimes hear it referred to as a MAC address. The term MAC address is especially relevant when we're talking about Ethernet networks. MAC addressing comes in several different types. First we have the basic MAC address type. The basic MAC address type is expressed in hexadecimal format as is all types of MAC addresses, and it is 48 bits long. However, even though the basic MAC address is 48 bits long, it actually has two parts.
The first part of the MAC address is called the Organizationally Unique Identifier or OUI. OUI is 24 bits long and it is very specific to the manufacturer of the device. Each manufacturer has its own 24 bit Organizationally Unique Identifier. In fact, the manufacturers need to register their Organizationally Unique Identifiers with a registration authority in order to make sure that no two manufacturers are using the same OUI.
The second part of the MAC address is something called the host portion of the MAC address. The host portion of a basic MAC address is also 24 bits long, however, unlike the Organizationally Unique Identifier portion, which is set, the host portion actually changes for each device created. This allows for over 16 million unique MAC addresses to be applied to every run of a device that a manufacturer makes before the manufacturer has to start reusing the host IDs again.
The reason it is important that a manufacturer minimize the reuse of the host portion of the MAC address is because no two devices on the same network can share the same MAC address. Whereas if I have two devices on a segment in my network and they both have the same MAC address, it will actually cause interference and problems on my network. And so with over 16 million unique host identifiers, it's very rare that you'll ever run across an error on your network, it does however happen once in a very great while.
Now besides the basic MAC address, there are actually a couple of variations that are widely used of the MAC address as well. The first variation is called the 60-bit Extended Unique Identifier or EUI-60. As the name implies, this MAC address is 60 bits long instead of 48 bits. However, the Organizationally Unique Identifier portion is still going to be 24 bits long and it still has to be the one that the manufacturer registered with the registration authority.
The host portion of this particular MAC address however is 36 bits long. This allows for even more than the 16 million unique MAC addresses that the original basic 48-bit MAC address allows. Aside from the length of the MAC address and the size of the host portion of that MAC address, the 60-bit Extended Unique Identifier MAC address works just the same as the conventional or basic MAC address we talked about previously. Another variation on the basic MAC address is the 64-bit Extended Unique Identifier or EUI-64.
Again, as the name implies, this is a 64-bit long MAC address. Again, the Organizationally Unique Identifier is still noted 24 bits and still has to be the 24-bit identifier that an organization registered with the registration authority. However, in the 64-bit EUI version of the MAC address, the host portion is up to 40 bits long. This provides for almost unlimited numbers of unique MAC addresses.
The 64-bit EUI MAC address can interact with an IPv6 IP address and together they can generate unique logical addresses that can be used on a network without running the risk of using the same logical address twice. This is very useful if you don't have a DHCP server or whatever but you're still running IPv6. This way you can still automatically generate unique logical addresses without having to worry about repeating them.
Now that we've talked a little bit about MAC addresses, lets go ahead and see how this works out in a diagram. What I have here is a diagram of a theoretical network. Right now this network has no address in it whatsoever. I'll now go ahead and place a unique MAC address on every interface in this network. You will notice that the computers and the round devices, which are routers, all have MAC addresses, but the square devices, which are switches, do not.
The reason the switches don't have MAC addresses is because effectively the port in the switch takes on the MAC address of whatever device is connected to it. The router on the other hand, because it's a unique network interface controller, has to have a unique MAC address just like the PCs on the network do. As a result, we have the network presented before us with physical addresses assigned to every interface on the network.
The routers will have multiple interfaces and therefore they have to have multiple MAC addresses because we cannot use the same MAC address twice on the same network. Once we know what the MAC address is for every device on our network, we can now begin to form our network into smaller segments. A segment is formed when one or more switches connect computers to a router's interface. Another way you can form a segment is when two routers are connected directly to each other.
Finally, segments can also be formed between just switches without a router being involved. However, if a router is not involved, that network will not be able to communicate or connect to an outside network. In this diagram here, each switch along with the computers and the router interface connected to it becomes a segment. Also the interfaces that connect the routers directly together to each other also become a segment. As a result, we have 10 different segments up here on this diagram.
- Binary, hexadecimal, and octal numbering
- How logical and physical addressing work together
- Comparing broadcast and collision domains
- IPv4 classless addressing
- Subnetting with IPv4 and IPv6
- IPv6 reserved ranges
- IPv6 link local, 6to4 tunneling, and 4to6
- Configuring DHCP and DNS in Windows Server
- WINS and NAT
- Port, packets, and remote access