A response may be sent in response to a request or as a part of a regular update. The request message may contain some or all of the complete routing table. Depending on the request packet, the response is sent with all or part of the routing table.
Version 1 is addressed for this discussion. Version 2 is addressed in Chapter 7, "Routing Information Protocol Version 2." (See also RFC 1723.) The address family identifier is currently only for IP, and the value of IP is two. None of the known IP implementation identifies any other address family. The IP address in the header is four octets in the network order. The metric field must contain values between one and 15, and this specifies the routers' current metric to reach the destination or determines that the value of 16 is unreachable. The maximum possible datagram should be 512 bytes, so by calculation, there could be 20*25 (number of routes) + 4 bytes for a common portion. This limits the update packet to fewer than 512 bytes.
To explain further: The first four bytes of the RIP header are common to every routing entry in a RIP packet. First, there is the command and version, and then the 20 bytes must change according to the route. There are two reasons for this: First, with every routing entry, a different IP address is advertised. Second, each IP address has a different metric.
Therefore, the equation becomes the following:
4 (common part) + [20 (bytes header changes with each entry)*25 (routes)] > 512 bytes
RIP does not support VLSM (variable-length subnet mask). This causes a lack of address space because, for serial lines that have point-to-point connections, smaller masks can cause address waste. Therefore, it is important to understand how to utilize RIP in a VLSM environment. For a description of subnetting, refer to Chapter 2, "IP Fundamentals." If you recall from basic VLSM, when a major network has different masks for different types of networks, perhaps to accommodate the number of host addresses, you can use a different subnet mask for broadcast networks and a different subnet mask for point-to-point networks. RIP does not support VLSM. However, to utilize RIP in a VLSM environment, consider the following courses of action.
Assume, for example, that a network is migrating from RIP to OSPF. On the OSPF routers, VLSM has been configured. To propagate these routes into RIP, its mask must be matched (see Figure 6-8). As mentioned previously, RIP does not propagate route information about a destination whose mask does not match that of its outgoing interface. For example, in Figure 6-8, router R1's serial link is connected to network 18.104.22.168, and the subnet mask of the interface subnet is 255.255.255.252. R1 also has an Ethernet connection to the same major network 22.214.171.124, and the subnet mask of Ethernet is 255.255.255.0. Because R1 is running OSPF on the serial line and still has legacy RIPV1, RIP must receive route information about the serial interface. By default, RIP will not work with VLSM behind the Ethernet of R1. When R1 attempts to send routing information about its serial interface from the Ethernet, it cannot do so because of the unmatched masks. Any subnetwork that does not have the same mask as that on the Ethernet will have connectivity difficulties. Therefore, those subnetworks will not be propagated in RIP domain.
As shown in Figure 6-8, R1 is the redistributing router used to propagate subnet 126.96.36.199 into the RIP domain. The mask of the subnet should match that of the outgoing interface of RIP. In this case, the Ethernet mask is (255.255.255.0) 24-bit, and the serial is (255.255.255.252) 30bit. You can create a route to subnet 188.8.131.52 that matches the RIP mask, and then redistribute that route into RIP. In this case, all the routers behind the Ethernet of R1 receive the route to destination 184.108.40.206 255.255.255.0. By default, RIP checks the interface mask on which it sends routing updates before actually sending these messages. When a route is informed about a destination that belongs to the same major network as the one on which updates are being sent, its mask is verified. If the mask of the connected interface is the same as that of the route being advertised for the same major network, then that route is advertised. Otherwise, the update for that subnet is dropped.
Then R1 would have two routes to 220.127.116.11—the first is the serial interface address of 18.104.22.168 255.255.255.252 (/30 mask); the other is 22.214.171.124 255.255.255.0 (/24 mask). When a packet is received that must be routed (such as 126.96.36.199, which is the IP address of the other end of the serial link), R1 compares its two routes. The first is /30 mask route and the second is the /24 mask route. The longest prefix mask wins—in this case, it is /30, and the routing continues, uninterrupted. The router, in this case, would select a 30-bit mask. When the router recognizes that a packet has been sent to a destination, the mask closest to the destination address packet is routed toward that address. For example, when router R1 in Figure 6-9 has a packet to send to 188.8.131.52/32, it has two choices: One is to send it to 184.108.40.206/30, and the other is to send it to 220.127.116.11/24. The longer the prefix mask, the more specific it is, making it more acceptable to the router..
Was this article helpful?