Day 23. IP Addressing

Notation Example: xxx.xxx.xxx.xxx

IP Address Example: 0-255.0-255.0-255.0-255

One of the reasons TCP/IP is used so prevalently is its capability to identify both the network and the host. Because an IP address does not always show which portion is the network and which is the host, a subnet mask is used to determine each. We cover subnet masks later. Every computer on a single LAN must have the same network address, and it also must have a unique host address. If two computers have the same host address, both will indicate an error message.

IPv4 uses 32-bit (4-byte) binary addresses, which limits the total number of possible IP addresses (or hosts) to 4,294,967,296. There are five classes of IP addresses: A–E. Classes A–C are used in networks, and Class D and E are reserved for multicast and research. Each class of addresses uses more of the address to determine the network portion and less to determine the host. Class A, for example, uses only the first 8 bits, or the first octet, for the network and the last three octets for the hosts. This gives a Class A network only 126 possible networks and 16,777,214 hosts. Table 23-1 provides a list of each address range and the possible number of networks and hosts.

Table 23-1 TCP/IP Address Ranges

To configure the IP address in Windows, go to the Control Panel and then the Network and Sharing Center; then select the link to the Local Area Connection. (In Vista, select the Manage My Network Connections link.) Under Properties, select IPv4 and then select its Properties option. The General tab contains both a method to set a static IP address and a method to obtain an IP address automatically using DHCP. See Figure 23-1 for an example of the Windows IPv4 configuration applet.

Figure 23-1 IPv4 Windows Configuration

Table 23-2 Default Subnet Masks

Another way to determine to which class an address belongs is to look at the first octet. The range column in Table 23-3 provides a breakdown for each class. For example, any address that starts with a 193.xxx.xxx.xxx will be a Class C address. Any address that starts with a 104 will be a Class A address and so on.

Table 23-3 IP Class Ranges and Usage

Table 23-4 Special IP Address Ranges

Table 23-5 Private IP Addresses

Figure 23-2 Host Addressing Guidelines Example

Host addressing guidelines:

The host portion identifies the device connecting to the network, and it must be unique to the network.

No address can be set to all 1s because this is a broadcast address (255).

No address can be set to 0 because this is reserved to identify the IP network ID.

When applying IP addresses in a network, remember the following rules about addressing in order to prevent making mistakes when addressing multiple networks. Figure 23-3 provides an example of three networks connecting back to a single network. Note how each WAN connection has its own unique network address and each LAN has its own unique network address.

Figure 23-3 Network Addressing Guidelines Example

Network addressing guidelines:

The network ID must be unique to the internetwork. All networks that connect to each other must be unique.

If an address is accessing the Internet, it must be unique from every other address. (In other words, it must be a public address controlled by IANA.)

Windows addressing guidelines:

Addressing a host can be done automatically through DHCP, which is called dynamic addressing. It also can be manually configured, which is called static addressing.

To prepare a computer for addressing, start by going to the Network and Sharing Center, then go to Properties, and select IPv4. Finally, look at the properties of the IPv4 selection. To access the local network, the IP address, and a subnet mask are the only items that need to be configured. If accessing the Internet, several pieces of information are needed besides the IP address: the subnet mask, a gateway address, and at least one DNS server address.

A router is used to connect one network with another network. Its job is to read the IP network address and determine whether traffic is meant to stay on the local network or needs to be forwarded to a port on the router so it can be sent to another port and out to another network. It uses tables to keep track of routes and network addresses.

Routers have a unique network address on each port. One port is usually an RJ-45 connection and faces the local network, while the other side can be a serial port that connects to the telecommunications company. The RJ-45 port on the router is usually considered the default gateway for computers on the inside network. Figure 23-4 provides an example of a router with both an Ethernet port and a serial port.

Figure 23-4 Router with Gateway Address

If a computer cannot obtain an IP address from a DHCP server, it will use an APIPA address. Although it can access the local network with this address, it cannot get to the Internet.

In CIDR, the idea is that any network address can use any subnet mask. For example, instead of a 10.0.0.1 address using a subnet mask of 255.0.0.0, which makes the 10 the network portion of the address and the 0.0.1 the host, using a Class C subnet mask of 255.255.255.0 changes the network and host separation. It means the 10.0.0 is the network and just the .1 is the host.

CIDR is written as a /nn with the number of bits borrowed from the host side of the address as the nn. For example, 192.168.1.1/24 refers to the fact that this is a Class C address and the first 24 bits represent the network (the 255.255.255 portion).

This isn’t the best example because a 192.168.1.1 with a subnet mask of 255.255.255.0 is still a classful address. So, let’s change our example to a Class A address of 10.10.1.1/16. Now we have a Class A address using a 16-bit subnet mask. Remember that each octet contains 8 bits, so 16 bits means we will be using only two sets of 255s. We have a Class B subnet mask of 255.255.0.0. Our Class A network address is now being treated as though it were a Class B address. Therefore, the new network portion of this address is 10.10 and the new host portion of the address would be 1.1.

The addressing format for IPv6 uses eight sets of 16-bit hexadecimal numbers separated by a colon (:). Because it is a hexadecimal number, letters A–F are used within the address. They are not case sensitive. Here’s an IPv6 IP address example:

2001:0db8:0000:5784:8af3:0000:0548:6354

A full IPv6 address comes in three parts: the global routing prefix, a subnet, and then the interface ID. For the address shown previously, the 2001:0db8:0000 is the global portion, 5784 is the subnet, and the rest is for the individual interface ID.

An IPv6 address can be written using an abbreviated method in the following way:

Leading zeroes can be omitted (this includes a set of four 0s being truncated to a single 0).

Consecutive zeros in a contiguous block can be represented by a double colon once in a single address.

For example:

2001:0db8:0000:0000::8af3:0000:0000:004b

Can be written as

2001:db8::8af3:0:0:4b

Unicast addressing refers to the method of sending messages between a single sender and a single receiver over a network. It must be unique on a network segment. To communicate, a Unicast message is sent to the nearest host closest to the sender. In other words, it works by sending packets to the nearest device in a group of previously configured receivers instead of a one-to-one, a one-to-many, or a one-to-all.

Anycast addresses are a type of addressing method that can be assigned to more than one device. Multiple hosts can be assigned the same Anycast address. This is unique to IPv6.

Multicast refers to the method of sending messages from one or more senders to a group of other specific receivers. All hosts must join the multicast group before receiving traffic.

IPv6 multicast addresses have an 8-bit prefix of FF00 with the second octet defining the lifetime and scope of the address. IPv6 has the capability to autoconfigure itself without the need for a DHCP server. With that being said, multicast addresses can be used for devices to acquire an IP addresses from a DHCPv6 server.

Use Table 23-6 to help memorize the IPv6 address types.

Table 23-6 IPv6 Address Types and Descriptions

Broadcast refers to the method of sending messages between a single sender and multiple receivers over an entire network segment. In IPv6, broadcasts are no longer used because they use so much bandwidth and are unnecessary.

Loopback addresses in IPv4 are represented by 127.0.0.1. In IPv6, the loopback address is 0:0:0:0:0:0:0:1. It can be written as ::1. A loopback address is used to test the local connection and cannot be assigned to a device or used to transmit data.

Link-Local addresses—Automatically assigned as soon as IPv6 is enabled, these are confined to a single link. They are identified by the first 10 bits (FE80::/10) and are not routable. Similar to a private IPv4 address.

Unique-Local addresses—Globally unique, but are used for local communication and always start with FD (FDE4:8db0:92e1::be).

Global Unicast addresses—Globally unique, which means they can be routed across the entire Internet. They are similar to an IPv4 public address. The prefix for a Global Unicast address is currently 2000::/3.