Variable-Length Subnetting in the RFCs
| Ordinarily, I would point out an IETF source document that is the basis for an Internet technology and then expound on that technology. In the case of Variable-Length Subnet Masking (VLSM), there is no clear-cut genesis document. Searching the Internet or the RFC Editor's database turns up a variety of references, mostly in documents dedicated to other topics. The more salient of these tangential reference documents are RFC 1009 and RFC 1878. They provide you with the context for the development of variable-length subnets and supporting mathematics and helps you appreciate a more thorough examination of VLSM. The following sections discuss each of these RFCs. RFC 1009The first "official" acknowledgment that you could use multiple subnet masks came in June 1987 with RFC 1009. Although this RFC focused on requirements for Internet gateways, the perceived benefits of a flexible approach to subnetting were identified. The rationale for supporting flexible or variable subnet masks was an acknowledgment of the inefficiency of trying to use a single mask for multiple subnets. Additionally, the RFC's authors acknowledged that subnetting was strictly a local phenomenon that had no impact on global routing. Consequently, enabling subnets to be created with different sized masks within the same network address offered tremendous benefits with no disadvantages. The flexibility was deemed critical to a continued ability to cope with the Internet's future expected growth. Here are some of the rules stipulated in this RFC:
Remember that because a subnet mask is of local significance only, nothing you do at the subnet level has any impact whatsoever on routing to your network. At the time RFC 1009 was published, many network administrators had figured out the mathematics of VLSM on their own and were not only creating variable-length subnets, but also nesting multiple levels of subnets within other subnets! Thus, the Spartan description of guidelines on how to support variable-length subnets in RFC 1009 amounted to little more than an acknowledgment of what was already becoming common practice. The term VLSM is not used in RFC 1009. Instead, the RFC's authors seem to prefer describing this phenomenon as flexible subnet masks.
RFC 1878RFC 1878 is an Informational RFC released in December 1995. It defined no new technology or protocol, but it offered greater clarification on the mathematical trade-offs between the number of hosts and the number of subnets that could be created with various-sized network blocks. In fact, RFC 1878 was titled "Variable Length Subnet Table for IPv4." One of the more useful tables in this RFC is excerpted in Table 4-1. This table demonstrates the correlation between the number of subnets and hosts you can define with any given-size mask. The mask size is indicated in both the familiar decimal terms and a new notation that was introduced in CIDR. This notation explicitly identifies the extended network prefix by using a slash (/) followed by a number. The slash can be thought of as a flag that indicates that the following numbers specify how many bits in the IPv4 address are used for network and subnetwork addresses. Thus, the number after the slash, when subtracted from 32, yields the number of bits allocated to host addressing. Please note that all values in Table 4-1 are gross and are not adjusted for usability.
Table 4-1 should exhibit some interesting mathematical patterns. If these patterns aren't yet self-evident, fear not: It gets easier with practice! Basically, you will see a repetition of a numeric sequence in the Decimal Mask column. A single bit set equal to 1 in the leftmost column of a binary octet carries a decimal value of 128. The initial 2 bits, when set to 11, yield a decimal equivalent of 192. Thus, the increments of decimal numbers must follow the pattern you first saw in Chapter 2: 128, 192, 224, 240, 248, 252, 254, and 255. The next interesting pattern you should recognize is that, starting with the /32 mask (which references just one host address), you are doubling the number of possible host addresses with each bit you add to the host field. This doubling is complemented with a halving of the number of subnets available. Start at the bottom of the Number of Possible Host Addresses column. For every bit you add to the host address field, you double the quantity of addresses available. With each bit you remove from the host field, you are halving the number of available hosts in that sized network. Thus, a /25 network offers exactly half the number of host addresses that are available in a /24 network. Understanding this relationship is necessary for understanding both VLSM and CIDR. We'll look at CIDR in much more detail in Chapter 6, "Classless Interdomain Routing (CIDR)." If you'd like to read more about the guidelines for VLSM contained in RFC 1878, here's the URL:
|
EAN: 2147483647
Pages: 118
- Designing MPLS Layer 3 Site-to-Site VPNs
- Comparing IPsec Remote Access VPNs with Other Types of Remote Access VPNs
- Using Clientless SSL Remote Access VPNs (WebVPN) on the Cisco VPN 3000 Concentrator
- Implementing Full Network Access Using the Cisco SSL VPN Client
- Strengthening SSL Remote Access VPNs Security by Implementing Cisco Secure Desktop