| Previous | Table of Contents | Next |
The IP protocol suffers from both the manual assignment noted previously and a shortage of legal addresses. Later in this chapter, one solution to this problem will be presented—the use of private addresses. However, conservation of address space can also become a concern with private addresses.
|
Network Design in the Real World: Addressing It would be hard to believe that a corporation with only a few hundred routers could use all of its addresses in a three-year time frame, but it does happen. The most significant contributor to the exhaustion of addresses is the lack of VLSM support. Being forced to use a consistent mask for all addresses quickly leads to hundreds of addresses being unallocated on point-to-point links and other small segments. One such network used all of its upper two private address spaces (RFC 1918 is defined later in the chapter) and all of its public Class C address blocks. When each of the few hundred routers contained at least three inter-faces, and many included 10 to 20, the addresses became exhausted. Secondaries and poor documentation further added to the problem. Ultimately, a complete readdressing strategy was needed, and encompassed in this plan was a change of routing protocol to support VLSM. This required a great deal of resources and a large expense—ideally, having a VLSM-aware protocol would have prevented the problem. You may point out that VLSM-aware protocols are relatively new and some of these networks are relatively old. That is true. And many of these networks needed additional addresses that were assigned via secondaries. This eventually led to bigger problems since troubleshooting and documentation were greatly affected. Today, no organization should continue to use secondaries and non-VLSM-aware protocols as a strategic direction. The penalties of not migrating in terms of hidden costs are too great to ignore in the long run. |
Table 3.2 documents the common subnet divisions used by network designers. It is important to note that 24- and 30-bit subnets are used most commonly—LANs using 24 bits and point-to-point WAN links using 30 bits. The number of subnets referenced in Table 3.2 presumes a Class B network—other base classes will differ.
| Number of Network Bits | Subnet Mask | Number of Subnets | Number of Hosts Per Subnet |
|---|---|---|---|
| 18 | 255.255.192.0 | 2 | 16,382 |
| 19 | 255.255.224.0 | 6 | 8,190 |
| 20 | 255.255.240.0 | 14 | 4,094 |
| 21 | 255.255.248.0 | 30 | 2,046 |
| 22 | 255.255.252.0 | 62 | 1,022 |
| 23 | 255.255.254.0 | 126 | 510 |
| 24 | 255.255.255.0 | 254 | 254 |
| 25 | 255.255.255.128 | 510 | 126 |
| 26 | 255.255.255.192 | 1,022 | 62 |
| 27 | 255.255.255.224 | 2,046 | 30 |
| 28 | 255.255.255.240 | 4,094 | 14 |
| 29 | 255.255.255.248 | 8,190 | 6 |
| 30 | 255.255.255.252 | 16,382 | 2 |
Designers should consider the following factors when allocating subnets:
| Network masks may be written in various formats. The mask 255.255.255.0 may be written as /24, to reflect the number of ones in the mask. | |
Today, network design requires a thorough understanding of TCP/IP addressing in order to be successful. Most of this requirement is facilitated by the explosive growth of the Internet (and its use of the IP protocol); however, the IP protocol also scales well, which generates benefits when it is used in the private network.
Unlike AppleTalk and IPX, IP addressing and routing benefits from summarization and other design criteria that are not available in the other protocols’ addressing schemas. IP permits efficient and logical addressing based on various criteria—unfortunately, most current networks evolved, rather than planned, their addressing schemes, effectively negating any benefits that may have been available from the protocol itself.
The design of IP addresses in the network requires the organization to make a number of decisions. These decisions concern:
| Previous | Table of Contents | Next |