Designing a Network and Routing Topology for a Company
| Routing is the process of using protocols, sets of rules, to determine the destination of a packet and deliver the packet safely to its destination. The protocols used on a network are part of its logical and physical design. Certainly the most prominent protocol today, by far, is TCP/IP. We therefore need to understand TCP/IP addressing methods and standards and their effect on our network design. In addition, we need to choose an addressing scheme that can include registered and/or private addresses. In this section, we examine the options available to us in regard to IP address assignment, which include
Static AssignmentYou can assign a static IP address by typing it in the Internet Protocol (TCP/IP) Properties dialog for the connection, as illustrated in Figure 6.5. Figure 6.5. You can set a static address for a TCP/IP host.
You can assign a static address to any host on the network, but static addresses are typically used for the following:
Let's now discuss each of these components and the reasons that a static IP address is the preferred option for each. ServersModern networks use many different types of servers. Many of these servers, such as DNS, DHCP, and WINS servers, require a static address configuration. Not all servers require a static address, but it is generally accepted as good practice to give all servers a static address whether or not they require it. Network PrintersNetwork printers are print devices that do not require a print server to function on the network. They have their own network interface card (NIC) and can be assigned an IP address. Because these devices are static and clients need to find them quickly and easily, it makes sense to assign them static addresses, rather than have the addresses change and possibly cause connection problems for the clients. As a side note, these devices can easily be assigned IP addresses from a DHCP server by using an address from a reserved pool that is permanently assigned to the devices' MAC address. Router InterfacesBecause router interfaces represent a fundamental part of the logical and physical network, they should always be assigned static addresses. Most routers don't have the capability to obtain an address automatically. Even if they do, you should take great care when considering this practice because router interfaces are configured on clients and servers. In other words, clients and servers rely on that router interface (default gateway) to be where they think it is! If it moves around, the whole network topology is disrupted. Automatic Private IP Address AssignmentAll Microsoft clients newer than Windows 95 OSR2 are set by default to obtain an IP address automatically as soon as they are installed and configured to use the network. There are many different ways in which they might obtain an address automatically. We discuss some of those methods later in this chapter. Microsoft clients, including Windows 98 and all newer clients, can also provide themselves with an automatic private IP address assignment (APIPA) if they find there are no network components available to provide them with a valid network address. Automatic private IP addresses are set by the client and are chosen from the address range of 169.254.0.1/16 to 169.254.255.254/16. The address will be chosen at random and then pinged to ensure that no other client on the local subnet is using it. This is a convenient service to connect clients when a DHCP server is not available or when a small workgroup does not require automatic address assignment by other means. However, this method has serious limitations. You should be aware that this service can allow clients to see each other, but it does not really connect any of the clients to the resources on the actual network. In other words, all clients that have addresses in this range can see each other, but none can see the network resources, which are on another subnet. Misery loves company! This could potentially create a troubleshooting nightmare for you if you don't immediately recognize the 169.254.0.0/16 network address. For this reason, some administrators choose to disable APIPA. Dynamic Host Configuration ProtocolThe easiest and most accurate way to assign addresses to clients in any network is dynamic host configuration protocol (DHCP). All clients set to obtain an IP address automatically will broadcast on the network to attempt to discover a DHCP server. When a DHCP server is available and properly configured, it responds to the client's request and assigns the client an IP address and potentially much more. DHCP servers use scopes to determine the range of IP addresses to assign. Scope options can include the address of other servers and connections in the network. For example, a DHCP server can assign the client an IP address as well as the IP addresses of a DNS server, WINS server, and default gateway, just to name a few. You can configure a DHCP server to automatically assign client addresses based on many factors regarding the client. Those factors could include the physical subnet in which the client resides, the type of operating system on the client, and even the media access control (MAC) address of the client. Figure 6.6 illustrates some of the options available on a DHCP server. Figure 6.6. You can configure many options on a DHCP server.
Types of IP AddressesThere are two main types of IP addresses in any network:
In this section, we examine each of these types of addresses and the relation of each to the physical design of our network. Public AddressesPublic IP addresses are real addresses that take up a part of the finite address space on the Internet. Therefore, they have to be assigned from the organizations that manage the Internet. One of these is the Internet Corporation for Assigned Names and Numbers (ICANN). It is a nonprofit corporation that was formed to assume responsibility for IP address space allocation and domain name management. Public IP addresses are registered through ICANN or other organizations that work through ICANN throughout the world. Other organizations include the InterNIC and the Internet Assigned Numbers Authority (IANA), which has its own department for each part of the world. In North America, we use the American Registry for Internet Numbers (ARIN). If all of this sounds confusing, that's because it is. But each entity has a Web site that can direct you to the correct site and assist you in registering IP addresses and domain names. All public addresses must be registered. Using our current version of IP addressing (Version 4), which uses a 32-bit address, we have only a little more than 4 billion public IP addresses for the entire world! In the early 1990s, this seemed like a gigantic number of addresses and there were no concerns of running out. Today, only 10 years later, we have about 1 billion public addresses remaining. Changes in the way network addresses are issued have slowed the rate at which we're consuming the public IP addresses, but we will still run out at some point in the future. A new version of IP addressing (Version 6) is on the horizon. This will use a 128-bit address and provide an enormous number of IP addresses. In fact, the number is so large that it is difficult to comprehend. The number of IP addresses that will be provided by IP Version 6 is 340,282,366,920,938,463,463,374,607,431,768,211,456 This will provide at least 1,564 addresses for each square meter of the surface of the planet Earth! Private IP AddressesThere is one more way that we can preserve our current IP address space. Using the technology that we have discussed in particular, NAT, firewalls, and proxy servers we can use significantly fewer public addresses. In a practical sense, we can use any address that we choose on the inside of our firewall, as long as that address is never seen on the outside. However, what if you and I both decide to use the same random address behind our firewalls, both of our firewalls fail, and our networks see each other? Then we would have an address conflict. To prevent this from happening, some address ranges are filtered by Internet routers, marking them as nonroutable, and are therefore recommended for use as private internal IP address ranges only. These address ranges include the following:
All the network addresses, broadcast addresses, and host addresses in these address ranges are filtered from entering the backbone routers of the Internet. You should use these address ranges for internal use on your network.
|
EAN: 2147483647
Pages: 152