Agreed Bob, I don't think a loop will form. If I understand correctly, EIGRP does not use split-horizon as a loop prevention tool, but rather as a way of reducing the size of the topology table. You can completely disable split-horizon in EIGRP, and DUAL will create a loop-free topology.
-------- Original Message -------- Subject: Re: [OSL | CCIE_RS] EIGRP in broadcast environment From: Bob McCouch <[email protected]> Date: Fri, August 10, 2012 2:08 pm To: Joe Sanchez <[email protected]> Cc: "[email protected]" <[email protected]>, Fulvio allegretti <[email protected]> Please elaborate. DUAL should ensure a loop-free topology by selecting the lowest metric path for each of R1 and R2. The path heard by the other router would be installed in the topology table, possibly as a feasible successor if the route metric is appropriate. Can you explain how the loop forms? Bob -- Sent from my iPhone, please excuse any typos. On Aug 10, 2012, at 2:00 PM, Joe Sanchez <[email protected]> wrote: I see a loop.. On Fri, Aug 10, 2012 at 8:20 AM, Bob McCouch <[email protected]> wrote: > Hi Fulvio, > > I think your confusion comes from the usual split-horizon definition > that "a route is never advertised out an interface it is learned > from," but that's an incomplete definition. Split-horizon comes into > play when the *best* route is learned from an interface, in which case > it will not be sent back out that same interface. Since R1 & R2 in > your example would each have their best path (successor route) through > their "other" interface, they would both advertise a path onto the > common segment for R1-4. R3 & R4 would learn equal-cost routes via > both 1 & 2. > > Hope that helps, > Bob > > Sent from my iPad > > On Aug 9, 2012, at 10:55 AM, Fulvio allegretti <[email protected]> > wrote: > > > Right, I know the best thing to do here would be to lab this and I will, > promise, but I'd like to know your understanding on this and if it is > something to worry about. > > Let's say we have 4 routers on a LAN all talking eigrp. Two of these > routers, router 1 and 2, have an eigrp path to subnet 1.1.1.0/24somewhere > else thorugh an interface not on the LAN and lets's assume for > sake of argument that the metric to 1.1.1.0/24 is the same for both of > them. Considering split-horizon there will be only one router, 1 or 2, > injecting that route on the LAN, right? One of the two routers will have > its best path to 1.1.1.0/24 through its own interface but won't inject it > on the lan as it heard it injected on the LAN from the other one. > > 1) What decides which one of the two will inject the route on LAN? Is it > the first router timewise that injects it on the LAN? > > 2) Let's assume R1 did, what happens if the metric through R2 > "improves", would the LAN routers choose the path through R2 now? > > Fulvio > > _______________________________________________ > > For more information regarding industry leading CCIE Lab training, > please visit www.ipexpert.com > > > > Are you a CCNP or CCIE and looking for a job? Check out > www.PlatinumPlacement.com > > > > http://onlinestudylist.com/mailman/listinfo/ccie_rs > _______________________________________________ > For more information regarding industry leading CCIE Lab training, please > visit www.ipexpert.com > > Are you a CCNP or CCIE and looking for a job? Check out > www.PlatinumPlacement.com > > http://onlinestudylist.com/mailman/listinfo/ccie_rs > _______________________________________________ For more information regarding industry leading CCIE Lab training, please visit www.ipexpert.com Are you a CCNP or CCIE and looking for a job? Check out www.PlatinumPlacement.com http://onlinestudylist.com/mailman/listinfo/ccie_rs _______________________________________________ For more information regarding industry leading CCIE Lab training, please visit www.ipexpert.com Are you a CCNP or CCIE and looking for a job? Check out www.PlatinumPlacement.com http://onlinestudylist.com/mailman/listinfo/ccie_rs
