TCPIP Configuration

TCP/IP Configuration

Of the three local area network protocols supported by Windows XP, TCP/IP requires the most configuration. At a minimum, every computer that runs TCP/ IP must have the following configuration settings:

• IP address. A unique 32-bit number represented as a set of four octet values that uniquely identifies the computer on a network

• Subnet mask. A 32-bit number represented as a set of four octets that is used to determined whether or not computers reside on the same network subnet

TCP/IP uses the subnet mask to determine where other computers with which it needs to communicate reside. Every computer on the same network must be assigned the same subnet mask, regardless of the subnet on which the computer resides. By examining the IP address and subnet mask assigned to a computer, TCP/IP can determine whether or not it resides on the local subnet. If the destination computer does not reside on the same subnet, TCP/IP forwards any data that it has for the destination computer to the assigned default gateway, which then routes the data to the appropriate network.

In order to route data packets to a computer located on another subnet or network, a default gateway IP address must be provided. Windows XP sends all network data not destined for a computer on the local subnet to the router or computer that is located at this IP address.

Administrators may need to supply the following TCP/IP configuration settings:

• DNS. A name resolution service used on the Internet and on most subnetted networks to provide IP address to computer name resolution.

• WINS. A name resolution service used on Microsoft networks to provide IP address to computer name resolution. WINS is supported to provide backward compatibility for applications that still require its services to function.

Given the amount of configuration that every Windows XP Professional computer that runs TCP/IP requires, it might seem that TCP/IP is not as good a choice for smaller networks that might benefit from NetBEUI's self-configuring protocol and for networks that don't require connection to the Internet. However, Microsoft provides an automated solution that greatly simplifies the administration of TCP/IP and makes it the best overall protocol for networks of all sizes.

Working with IP Addresses

Normally, most computers have a single network adapter. Every computer on a network that uses TCP/IP must have a unique IP address for each network adapter that is installed. TCP/IP addresses must be unique, meaning that the same IP address cannot be assigned to two computers on the same network. If two computers are assigned the same IP address, a conflict occurs, and one or both of the computers may not be able to access the network.

IP addresses are made up of two pieces of information, a network ID and a host ID. The network ID identifies the network to which the computer is connected. The host ID provides a unique identifier for each computer on that network. An IP address is made up of 32 bits, as demonstrated below.

 10101001.11111110.11001100.00000011

IP addresses are organized into four parts called octets, each of which is 8 bits long. Each octet has a possible value of 00000000–11111111. People find that working with and remembering binary numbers is difficult. Therefore, TCP/IP allows for the use of decimal numbers, which are then automatically translated into binary as needed. For example, the following IP address is the decimal equivalent of the previous binary IP address.

 169.254.204.003

Tip 

Use the Calculator application provided by Windows XP to manually translate between binary and decimal IP addresses. On Windows XP, the Calculator application can be started by clicking on Start/All Programs/Accessories and then selecting Calculator. By default, the Calculator application only works with decimal numbers. Click the View Menu's Scientific option to expand the Calculator to display options for working with binary numbers.

A centralized International body known as the InterNIC manages IP addresses. The InterNIC sells blocks of these IP addresses to large ISPs and telecommunications companies, which then lease them to businesses and consumers. The InterNIC organizes all IP addresses into classes. The first three classes contain IP addresses assigned to businesses and consumers. These classes are shown in Table 16.1.

Table 16.1: IP Address Class Organization

Class	Network and Host IDs	Number of Networks	Hosts per Network
Class A	nnn.hhh.hhh.hhh	126	16,777,214
Class B	nnn.nnn.hhh.hhh	16,384	65,534
Class C	nnn.nnn.nnn.hhh	2,097,152	254

In a Class A network the first 8 bits are reserved for defining network IDs.There is a total of 126 Class A networks, each of which contains over 16 million IP addresses. Similarly, a Class B network uses the first 16 bits to define network addresses and the last 16 bits to define host addresses. There are over 2 million Class C networks, each of which has 254 IP addresses.

The first octet in an IP address can be used to determine its class. For example, the first octet in a Class A network will always be in the range of 1–126. Table 16.2 outlines the network range for each class.

Table 16.2: Network Identification

Class	Network
Class A	1–126
Class B	128–191
Class C	192–223

Note 

The network address of 127 is omitted from Table 16.2 because it is a reserved network address used by computers running TCP/IP to perform a diagnostic loopback test and cannot be used on networks.

Unfortunately, as the popularity of TCP/IP and the Internet has grown in recent years, the available supply of IP addresses has been used up. Fortunately, new techniques for distributing IP address assignments have evolved that allow classes of IP addresses to be combined (supernetting) or divided (subnetting). In addition, a new version of TCP/IP called Ipv6 has been designed that, when finally implemented, will greatly expand the available range of IP addresses.

Deciphering the Subnet Mask

In order for the operating system to be able to determine whether computers reside on the same network or on different networks, it must examine the computer's assigned subnet mask. Like the IP address, a subnet mask is a 32-bit binary number. Every computer on the network must have an IP address and a subnet mask. In addition, every computer on the same network must have the same subnet mask. Table 16.3 lists the subnet masks associated with the first three IP classes.

Table 16.3: Subnet Masks

Class	Network
Class A	255.0.0.0
Class B	255.255.0.0
Class C	255.255.255.0

As Table 16.3 shows, the subnet mask defined to a computer on a Class C network is 255.255.255.0.

Public and Private IP Addresses

One technique used by organizations that require Internet access is the use of private IP addresses. Using private IP addressing, the company only needs to lease a single public IP address through which all the company's Internet traffic will flow. This public IP address is then assigned to a computer or router that is connected to both the Internet and the company's private network. Since none of the computers within the company directly connects to the Internet, they can be assigned private IP addresses. The device that functions as the Internet gateway automatically translates requests for Internet access by hiding private IP address assignments from the Internet and presenting only its public IP address.

A private IP address is one that does not connect to the Internet and is only known on the network to which it is defined. Several ranges of private IP addresses have been reserved for this purpose. These IP addresses are outlined in Table 16.4 and will not be found in use anywhere on the Internet.

Table 16.4: Private IP Addresses

Network ID	Subnet Mask	IP Addresses
10.0.0.0	255.0.0.0	10.0.0.1–10.255.255.254
169.254.0.0	255.255.0.0	169.254.0.1–169.254.255.254
192.168.0.0	255.255.255.0	192.168.0.1–192.168.255.254

Whether to lease one or more public IP addresses and which set of private IP addresses to use, if any, are decided upon during the initial setup of a local area network. Once established, the administrator needs to work closely with network engineers to ensure that the IP address assignments are performed in sync with the overall organization of the network.

Subnetting

One problem with modern networks is that they are very chatty, meaning that even when network computers have no data to transmit, they still generate network traffic. As more and more computers are added to the network, more data traffic occurs, eventually slowing down the network. To keep network communications running smoothly, large networks are broken down into manageable smaller networks called subnets. This reduces the number of computers transmitting on any particular network segment at a time.

Network subnets are connected together using devices called routers. A router is a device that is connected to two or more network subnets and routes network data from segment to segment as required. Any traffic destined only for a local subnet is not allowed through the router to the other subnets, thus localizing as much network traffic as possible.

While subnetting minimizes network traffic and allows for the establishment of large enterprise networks, it also adds to the complexity of those networks. Specifically, it requires that a network's IP address assignment be modified in order to support subnetting (that is, the subdividing of a network into two or more logical subnets). Subnetting requires that one or more bits be borrowed from the host ID portion of the IP address so that they can be used in the network ID portion in order to define multiple subnets.

Typically, network engineers and architects are responsible for the overall design of the network and determine the network classification that is used and how it is then broken down into subnets. This includes defining the number of subnets and their range of available IP addresses.

Most networks use a TCP/IP service known as DHCP to store and manage this information. DHCP also manages the automatic assignment of TCP/IP configuration settings to network computers, alleviating the computer administrator of the responsibility of manually assigning TCP/IP settings. However, some environments choose to implement static IP address assignment, in which case the computer administrator is responsible for configuring TCP/IP settings. In order to perform this role, the administrator must work closely with network engineers and architects and assign IP address settings according to their overall design and specification.