Operation of RIPng

RIPng for IPv6 is based on RIPv2, but it is not an extension of RIPv2; it is an entirely separate protocol. RIPng does not support IPv4, so to use RIP to route both IPv4 and IPv6, you must run RIPv1 or v2 for IPv4 and RIPng for IPv6.

RIPng uses the same timers, procedures, and message types as RIPv2. For example, like RIPv2 it uses a 30-second update timer jittered to prevent message synchronization, a 180-second timeout period, a 120-second garbage-collection timer, and a 180-second holddown timer. It uses the same hop-count metric, with 16 indicating an unreachable value. And, it uses Request and Response messages in the same way that RIPv2 does. And, like RIPv2, Request and Response messages are multicast with the same few unicast exceptions that RIPv1 and v2 use. The IPv6 multicast address used by RIPng is FF02::9.

An exception to these parallel functions is authentication. RIPng does not have an authentication mechanism of its own, but instead relies on the authentication features built into IPv6.

There is, of course, no need for the RIPv1 compatibility switches used by RIPv2, because there is no backward support for IPv4.

Figure 6-10 shows the RIPng message format. Unlike RIPv1 and v2, which run at UDP port 520, RIPng sends and receives its messages at UDP port 521. Also unlike RIPv1 and v2, there is no set message size. The message size is dependent only on the MTU of the link on which it is being sent.

Figure 6-10. The RIPng message format.

• Command is always set to 1, indicating a Request message, or to 2, indicating a Response message.

• Version is always set to 1, that is, the current version of RIPng in RIPngv1.

• IPv6 Prefix is the 128-bit IPv6 destination prefix of the route entry.

• Route Tag is used in the same way the 16-bit RIPv2 Route Tag field is used: for transporting external route attributes across the RIP domain.

• Prefix Length is an 8-bit field specifying, in bits, the significant part of the address in the IPv6 Prefix field. For example, if the advertised prefix is 3ffe:2100:1201::/48, the value in the Prefix Length field is 48 (0x30). If the advertised route entry is a default route, the IPv6 prefix is 0:0:0:0:0:0:0:0 and the Prefix Length is 0.

• Metric is the same hop count metric used by RIPv1 and v2. But as the maximum possible value is 16, the field is reduced to 8 bits from the unnecessarily large 16 bits of RIPv1 and v2.

RIPng specifies a next-hop address the same way that RIPv2 does. That is, a valid non-zero next-hop address specifies a next-hop router other than the originator of the Response message and a next-hop address of 0:0:0:0:0:0:0:0 specifies the originator of the Response message. But rather than associate a next-hop address with each route entry as RIPv2 does, RIPng specifies the next-hop address in a special route entry and then groups all route entries that use the next-hop address after it. That is, the next-hop address specified in a next-hop route entry applies to all of the route entries following it, either to the end of the Response message or until another special next-hop route entry is found.

Figure 6-11 shows the format of the next-hop route entry. The 128-bit address is either the IPv6 address of another router or, if it is ::, specifies the address of the originator. The Route Tag and Prefix Length fields are set to all zeroes. Receiving routers recognize the next-hop route entry because its Metric field is set to all ones (0xFF).

Figure 6-11. A metric value of all ones specifies that the route entry is a next-hop route entry. All route entries following this one, until the end of the message or until another next-hop route entry is encountered, use this next-hop address.