follow-up based on an offline conversation I've been having with someone. parts edited because I don't know if the sender would want his name associated with my ignorance. ---------------------------------------------
well, you know, you're probably right. the receiving end would be the only place where the decision could be made as to how to interpret what is in the network field. interface = 22.1.1.5/24 other routers have interface with: -------------------------------------------- 22.17.1.4.24 193.1.1.4/32 22.1.1.3/32 network field routing table contains interprets ------------------- --------------------- 22.17.1.? 22.17.1.0/24 193.1.1.1 193.1.1.0/24 22.1.1.3 22.1.1.3 Us poor folk don't have sniffers to verify, but I am guessing that on the sending side, the RIP process places a value that makes sense based on the address and mask of the interface. 22.17.1.0 for example. Then on the receiving side, the RIP process looks at the value, and interprets it based on the mask of the matching interface if the value falls within the classful boundary, or summarizes at the classful boundary if the value in the network field does not fall within the classful boundary. So I get what you're saying. I have never really thought about it, concentrating rather on the desired result rather than the actual operation. Chuck Sent: Saturday, 11 January, 2003 11:26 PM Subject: Re: RIP1 > Thanks for the additional info. > > My point of view was inspired by the lack of a place in the packet to specify host, the lack of a method to distinguish a host in the spec itself, and the fact that inference of proper class is made by the receiving end in the customary fashion (them first three bits of the leftmost decimal octet). What this output further leads me to conjecture is that it's operating on the rightmost decimal octet to distinguish between hosts and networks, and that the work is still being done by the receiver. > > I only raised the point because the original question specified both advertising and receiving rip and separated the two cases. > > Chuck Larrieu wrote: > > interesting way of putting it. I can only prove the point by reverse logic. > > > > router 5 receives the following routes, as shown by the debug: > > > > 02:41:17: RT: network 22.0.0.0 is now variably masked > > 02:41:17: RT: add 22.2.2.4/32 via 22.1.1.4, rip metric [120/1] > > 02:41:17: RT: add 22.2.2.44/32 via 22.1.1.4, rip metric [120/1] > > 02:41:17: RT: add 197.1.3.0/24 via 22.1.1.4, rip metric [120/1] > > 02:41:17: RT: add 197.1.5.0/24 via 22.1.1.4, rip metric [120/1] > > 02:41:17: RT: add 22.2.2.3/32 via 22.1.1.3, rip metric [120/1] > > > > note the /32's > > > > this implies that the other two routers in the mix are sending /32's > > > > from the RFC: > > > > The RIP packet formats do not distinguish among various types of > > address. Fields that are labeled "address" can contain any of the > > following: > > > > host address > > subnet number > > network number > > 0, indicating a default route > > > > I believe the playpen now accommodates five numbers. wonder how long before > > it accommodates 6? > > > > Sent: Saturday, 11 January, 2003 10:22 PM > > Subject: Re: RIP1 > > > > > > > Correct me if I'm wrong, but that's achieved on ingress only, correct? > > > > > > If so, only half of what he's asking is possible (not that the other half > > matters, except for elusive goals such as understanding and absurd > > environments such as your favorite 4-figure playpen). > > > Chuck Larrieu wrote: > > > ""bergenpeak"" wrote in message > > > news:[EMAIL PROTECTED]... > > > > Is it possible, using RIPv1, to send advertisements which will > > > > be interpreted as /32s? I would think this is not possible > > > > as the route would be either advertised as a classful route > > > > (when crossing classful boundaries) or would be interpreted as a /30 > > > > or larger (based on how the receiving interface is configured). > > > > > > > > Is there some way to actually cause /32 routes to be advertised > > > > and interpreted as /32s in RIPv1? > > > > > > > > > as promised, from a real routing table. note that because of the classful > > > nature of RIPv1, the host routes must fall within the major classfull > > > network of the particular interface. Otherwise, what is received is a > > > classfull summary. For this example, note the /32's, indicative of a host > > > route. > > > > > > Gateway of last resort is not set > > > > > > > > > > -- ""bergenpeak"" wrote in message news:[EMAIL PROTECTED]... > Is it possible, using RIPv1, to send advertisements which will > be interpreted as /32s? I would think this is not possible > as the route would be either advertised as a classful route > (when crossing classful boundaries) or would be interpreted as a /30 > or larger (based on how the receiving interface is configured). > > Is there some way to actually cause /32 routes to be advertised > and interpreted as /32s in RIPv1? > > Thanks Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=60916&t=60885 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]