At 12:51 AM 2/22/02, Chuck wrote: >it gets complicated, routing protocols versus ip packets.
It's not complicated. Is this Chuck Larrieu? You know that it's not complicated. Of course IP routing protocol packets are carried in an IP packet. No biggie. They can set the IP TTL in the IP header to whatever they want. Routing protocols generally set the IP TTL to 1 in their routing protocol packets. That works fine because the recipient is next door. Note that we are not talking about the hop count in distance vector route descriptors carried by RIP, IGRP, and EIGRP. OSPF sets the TTL to 1. OSPF virtual links are a special case. The packet might have to go more hops, as you say. BGP also sets the TTL to 1. eBGP multihop might be another special case. RIPv1 and v2, IGRP, and EIGRP set the TTL to 2. Maybe the developers were worried that the recipient would decrement by one and trash the packet. Priscilla >first of all, if I understand correctly, all ip routing protocols use ip >headers. The routing protocol packet is the payload, and not an entity unto >itself. I have seen traces of OSPF packets showing IP TTL of various values. >Someone shared with me some traces to validate something I suspected - that >the OSPF virtual link packet has an initial TTL of 255. My theory is that it >has to be deliberately set high because there is no predicting the number of >hops a virtual link will traverse. > >The eBGP multihop command sets the IP TTL to something greater than the >native BGP TTL of 1. > >EIGRP? Don't know. Was merely speculating. But consider - where else might >the "hop limit" occur? The EIGRP header has no field indicating hop count >that I can see. My source is the Rad Com World of Protocols book. > >Yes, RIP and RIPv2 contain within the RIP packet ( not the IP header ) a >field in which metric / hopcount is carried. This leads me to believe that >RIP does nothing to manipulate the IP TTL value. The others appear to do >just that, however. > >Chuck > > >""Priscilla Oppenheimer"" wrote in message >[EMAIL PROTECTED]">news:[EMAIL PROTECTED]... > > At 08:05 PM 2/21/02, Chuck wrote: > > >to augment the other answers, the IP hop count is really the IP TTL >value. > > >It can never exceed 255 > > > > You're confusing two issues. > > > > Remember the router has two jobs: forwarding packets and learning the > > topology. Hop count has to do with the latter and affects what goes in the > > routing table. The IP TTL causes a router to drop a packet before > > forwarding if the TTL becomes zero. > > > > > > >EIGRP defaults to 100 hops, so I would expect that the routing packet IP >TTL > > >is set at 100 at that point. > > > > Routing packets only go to neighbors. The IP TTL should be set to one or > > two. This has nothing to do with hop count which will be later in the > > packet in the distance vectors. > > > > > > >Well ( checking the sniffer trace that Priscilla so thoughtfully supplied >a > > >couple of days ago ) I'm seeing the IP TTL as 2. Still, maybe there is an > > >adjustment made. After all, the (E)IGRP metric includes end to end >metrics. > > >hhmmm... ( looking over Priscilla's trace again ) way down there I see an > > >EIGRP hop count 0 line. > > > > The router was advertising a directly-connected network. > > > > > > >the IP TTL is still really the only thing that makes sense in terms of >the > > >way IP works. > > > > In terms of forwarding maybe. You better reconsider routing protocols > > though... > > > > Priscilla > > > > > > >Anyone? > > > > > >Chuck > > > > > >""Steven A. Ridder"" wrote in message > > >[EMAIL PROTECTED]">news:[EMAIL PROTECTED]... > > > > Anyone know why there is a hop-count in EIGRP? It has a 1 byte value, > > but > > > > it doesn't limit the number of hops and it looks like routers don't >use > > it > > > > in their calculations. Why is it there? > > > > > > > > -- > > > > RFC 1149 Compliant. > > ________________________ > > > > Priscilla Oppenheimer > > http://www.priscilla.com ________________________ Priscilla Oppenheimer http://www.priscilla.com Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=36247&t=36082 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
