Windows Server 2008 R2 networking : Planning and Deploying a TCP/IP Network Infrastructure (part 1)

Window networks depend upon a reliable TCP/IP infrastructure. A properly designed and managed TCP/IP network helps to ensure a successful Windows Server 2008 R2 deployment, while a poorly designed network almost guarantees that problems are going to occur during and after your deployment. Spend time to make sure that your network is healthy before rolling out Windows Server 2008 R2. If you already have a well-managed and reliable IP network, give yourself a pat on the back. This is not always an easy objective to accomplish.

Introduction to TCP/IP

Most of today’s networks, including the Internet, rely heavily on the TCP/IP protocol. The TCP/IP protocol stack has been around since the early days of computer networks and remains the de facto standard of enterprises today. Before setting up or managing a Windows network, you need to have a good understanding of how TCP/IP works. In this section, we will cover some of the basics of TCP/IP and how they apply to Windows. If you are already an experienced network administrator, now might be a good time to review and refresh your IP knowledge.

IP addresses

IP addresses are unique binary numbers assigned to hosts on an IP network. Think of IP addressing in the same way as you think of the addresses of houses in your neighborhood. Each house requires a unique street address. When someone needs to visit your home, they direct their vehicle to your address. The same applies in the world of TCP/IP networks. Every computer and device attached to the network requires a unique IP address. Data that needs to reach a certain computer on the network is sent to its IP address.

As mentioned, IP addresses are binary numbers; however, most people prefer to read IP addresses in decimal format for ease of use. It is important that you as a network administrator understand this concept to properly troubleshoot and manage IP networks.

IP address classes

IP addresses are distributed into five classes: Class A, Class B, Class C, Class D, and Class E. All IP addresses belong to a class based upon their decimal value of the first octet. Classes A, B, and C are the ones you will see used on corporate networks. Class D IPs are reserved multicast addresses that cannot be assigned to a single computer but used to send and receive multicast traffic. Class E addresses are reserved for use by the Internet Engineering Task Force (IETF). The IP classes and their corresponding range of IP addresses are listed in Table 1.

Table 1. IP Address Classes

Class	Range
A	1.0.0.1–127.255.255.255
B	128.0.0.1–191.255.255.255
C	192.0.0.1–223.255.255.255
D	224.0.0.1–239.255.255.255
E	240.0.0.1–255.255.255.255

IP subnetting

A subnet mask is another group of dotted decimal numbers, representing a binary number that distinguishes which part of the IP address represents the network. The subnet mask is used to allow computers to determine whether the addresses of other computers they wish to communicate with are on the local network or on a remote network. If the computer resides on a remote network, the communication request is sent to the default gateway. Figure 1 explains how subnet masks work.

Figure 1. How the Subnet Mask Works.

The three main IP address classes have default subnet masks. The standard subnet masks for each class, including number of supported hosts on each network are listed in Table 2.

Table 2. Standard Subnet Masks

	Subnet Mask	Number of Supported Hosts per Network
Class A	255.0.0.0	Over 16 million
Class B	255.255.0.0	Over 16 thousand
Class C	255.255.255.0	254

The default subnet mask is not practical in most network configurations. For example, let us say that you owned a Class B network of 159.247.0.0. Using the default mask, you could have over 16,000 computers on one nonroutable network segment. What if you had a remote office connected via a WAN link? Would you need to acquire another Class B network range for that office? First, this would be a major waste of your IP addresses and second, good luck on getting someone to give you that many. Luckily, you can create custom subnet masks to split up your IP addresses. By simply changing the subnet mask from 255.255.0.0 to 255.255.255.0, you have instantly given yourself 254 unique routable networks that can support 254 hosts each. Creating a custom subnet mask is as simple as adding some binary ones to replace zeros in the mask. But what if you need to support 400 computers in a remote network? What does the mask look like then? This is where it gets a little tricky. You will need to convert the dotted decimal to its binary equivalent and perform a simple calculation. Let us take a look at this process.

1. Decide how many subnets or networks you need to support. This is pretty easy to calculate. Figure out how many networks you have that are separated by a router.

2. Decide how many hosts you need on each network. You need to plan for the number of computers and other IP devices that you will want to support at each network location. Remember that you may need IP addresses for network switches, printers, and other IP-enabled devices on top of the number of computers that you need to support each network. You should plan for growth here as well. Give yourself at least 10% growth room for a given network.

3. Calculate the subnet mask. You now have enough information to calculate the proper custom subnet mask. Perform the following to calculate your subnet mask.

   1. Convert the standard subnet mask to binary. If we are using an IP network of 160.240.0.0, then the mask would be 255.255.0.0. The binary conversion is 11111111.11111111.00000000.00000000. Notice that it takes eight binary numbers to make up the number between each decimal. This is why each number between the decimal is referred to as an octet.

   2. Add one to the number of networks (subnets) you need. Assume that you need five networks. Add one to it to get six.

   3. Convert the decimal number to binary. You can do this manually or the calculator in Windows works great for this. In our example, we convert the decimal number six to binary, which is 110.

   4. Calculate the bits required for the mask. This is equal to the bits required to create the binary number. Since 110 is three individual numbers, 3 bits are required.

   5. Add the bits to the standard subnet mask resulting in a new binary subnet mask of 11111111.11111111.11100000.00000000. Now convert this binary back to decimal resulting in 255.255.224.0. You now have the subnet mask to use on each network segment.

Now that you have learned how to create a custom subnet mask, you should be aware that you can use a special subnet calculator to perform these steps for you. However, it is important that you understand how subnetting works if you plan on supporting Windows networks.

Public- versus private-IP addresses

IP networks expanded and grew much larger than the original creators of the protocol ever intended. IP blocks or classes were originally developed with a limited number of available addresses. With the emergence of global interconnected networks and the Internet, many organizations found themselves in an IP address shortage crisis. This is where private-IP addresses come into play. Private-IPs constitute a set of three IP address ranges, one from each of the three primary classes that are not routable on the Internet (see Table 3). The result of not making them Internet routable is that anyone can use them on their networks. If the private-IP addresses need to connect to the Internet, a Network Address Translator (NAT) device must be used to translate the private-IP to a public-IP. This technology allows organizations to purchase a limited number of public-IP addresses and use private-IP addresses on computers connected to their internal networks. The private-IP addressed computers can then use the NAT device, which is assigned a public-IP, to communicate on the Internet. A simple private-IP addressed network is depicted in Figure 2. The use of private-IPs and NAT not only decreases the usage of public-IP addresses, but also makes networks more secure by hiding computers from the global Internet. Private-IP addressing is a technology that continues to be available in IPv6.

Table 3. Private–IP Ranges

	IP Range
Class A	10.0.0.0–10.255.255.255
Class B	172.16.0.0–172.31.255.266
Class C	192.168.0.0–192.168.255.255

Figure 2. A Private-IP Network.

Introduction to IPv6

IPv6 is the next generation IP network protocol developed to replace the aging IPv4. As mentioned earlier, the designers of IPv4 never expected that billions of IP addresses would be needed to support the global networks we have today. Even with the increased use of private-IP ranges and technologies, such as NAT, the number of available public-IP addresses continues to decline. It has become very clear that future IP networks will require a lot more addresses than that are available in IPv4. This is where IPv6 comes in. IPv6 moves from 32-bit (4 octets) IP addresses to 128-bit IP addresses. This increases the number of available addresses to such a large number that every person on earth could have roughly 39614081257132168796771975168 addresses. Yes, that is a lot of IP addresses. The intent of the Internet Engineering Task Force (IETF), the governing body of IP networking, was not just to create some insanely large number just to ensure that we do not run out, but for easier management and assignment of IP ranges. IPv6 allows large blocks to be assigned, providing more efficient routing and easier administration of those IP ranges.

Though IPv6 is clearly the future of IP networks, the adoption rate has been very low to date. Major changes to enterprise networks, such as changing IP addresses, are never cheap or quickly implemented. Chances are that IPv6-based networks will emerge and grow over the next few years, but IPv4 will not be going away in the near future.

As a Windows administrator, the important thing to understand is that Windows Server 2008 R2 fully supports IPv6, and can efficiently communicate on both IPv4- and IPv6-based networks.

IPv4 to IPv6 transition technologies

To help organizations move to IPv6, there are several standards-based technologies that have been created to allow IPv6 applications function over an IPv4 network. Windows Server 2008 R2 includes support for some of these technologies, including Teredo, 6to4, IP-HTTPS, and ISATAP. A brief explanation of each is provided below:

• Teredo —Teredo is a standards-based protocol that provides IPv6 connections for IPv4-based computers that are behind an IPv4-based NAT. Teredo is a key technology allowing organizations to make IPv6 connections without changing IP addresses of computers on their internal private subnets.

• 6to4 —6to4 is a standards-based protocol that allows computers with public-IPv4 addresses to make IPv6-based connections over the IPv4-based Internet. It is a key technology allowing organizations to begin transitioning to IPv6 while the Internet at large continues to be based on IPv4.

• IP-HTTPS —IP-HTTPS is a Microsoft technology that allows Windows 7 and Windows Server 2008 R2 computers behind a firewall to establish IPv6 connectivity over an IPv4 network by creating an IPv4-based tunnel in which IPv6 packets can travel.

• ISATAP —ISATAP is a standards-based technology that provides IPv6 connectivity across an IPv4-based internal network.

Designing IP networks

The same requirement applies to building IP-based networks. You need to spend ample time planning prior to building your network infrastructure. Be sure to document your game plan so that you will not forget the critical tasks. Remember that your Windows network is not worth much if your workstations and servers cannot communicate. As part of your design, you need to understand and document what network services and applications you plan to support. You need to know how they communicate, what protocols they use, and how much bandwidth they require. You will also want to consider the following while developing your plan:

• Number of physical locations and logical networks

• Number of networks’ devices you plan to support

• Expected growth of your network

• Availability and redundancy requirements

• Bandwidth needs

• Routing options

• Network switch needs

• VLANs

• Network locations that will host servers

• VPNs and Remote Access technologies

• Internet access and firewall locations

These are just a few of the topics that you will need to spend time designing and documenting prior to deployment of an IP infrastructure. The end design should match up with your Windows Server 2008 R2 deployment plan. The IP infrastructure must be designed to support the various requirements of network applications provided by Windows Servers. In the end-user’s eyes, if the network is down, so are the services it supports.