Ok, I see, so if through the internal alogrithm iBGP wins, because of local pref for example, then the iBGP will be offered to the IP routing table and the EIGRP would be selected in your example. Great stuff. Thanks
From: [email protected] Date: Wed, 2 May 2012 14:16:17 -0400 Subject: Re: [OSL | CCIE_RS] iBgp vs eBgp path selection To: [email protected] CC: [email protected] Hi Fulvio, Think of it the other direction. All the protocols converge internally and use their internal algorithms to select its own best path to "offer" to the IP routing table. Among all routes offered to the IP routing table, the "best" one is then selected based on the AD of the protocol. In your example below, if the prefix is known via EIGRP, EBGP and IBGP, EIGRP will offer a route for the IP routing table at an AD of 90. BGP will go through it's process and if every other value is the same, eventually the external path will be chosen over the internal one, which will then be offered to the IP routing table. At that point, the IP routing selection process will pick the BGP route (which happens to be an EBGP route in this case, with an AD of 20) over the EIGRP route. If the route was only known via an IBGP path, it would be offered to the IP routing table with an AD of 200, and then EIGRP would win instead (and BGP would record a "RIB failure" for the route because it wasn't selected). Make sense? On Wed, May 2, 2012 at 6:43 AM, Fulvio allegretti <[email protected]> wrote: Thank Bob, OK that makes sense. I am familiar with BGP Best Path selection algorithm, but I thought it wouldn't even start the algorithm as AD are different, hence my second part of the question. But, as you say, admin distance is only taken into account to choose which routing protocol to use in the first place, then its down to the protocol algorithm. So, if we had a prefix known by igp, ebgp and ibgp and assuming default admin distance, the ebgp admin distance would make BGP the best routing protocol, but the path selection alogorithm could select the ibgp path due to a different parameter, local pref for example. How interesting. Thanks again Bob From: [email protected] Date: Tue, 1 May 2012 21:27:45 -0400 Subject: Re: [OSL | CCIE_RS] iBgp vs eBgp path selection To: [email protected] Hi Fulvio, This matter took me a long time to really understand. Administrative distance is used only to select the best route among multiple protocols for installation into the routing table. Each protocol has it's own selection algorithm to identify the candidate route it is offering for selection by the IP routing table, at which point AD is used to select the best overall route. In other words, Admin Distance cannot be used to select a specific route *within* a protocol, only among all protocols that have the route. For example, EIGRP will always select Internal routes over External routes. Even if you set the AD for external routes to 1, if the best route to a destination prefix is learned as an EIGRP internal route, that one will be installed with the AD of 90. AD doesn't matter when selecting *which* EIGRP route will be used, only when selecting EIGRP vs. another protocol. Likewise, OSPF will always select O, the O IA, then E1/N1, then E2/N2. Can't be changed (as far as I know). But then whichever route is selected will be installed in the IP routing table with a (default) AD of 110. For BGP, consult this: http://www.cisco.com/en/US/tech/tk365/technologies_tech_note09186a0080094431.shtml See Step 7. BGP External routes will always be chosen over Internal, all other values in steps 1-6 being equal. Unlike most other protocols, BGP has knobs that can be used to tune route selection before getting to Step 7. That's where weight, local preference, etc. come in. It took me a long time to wrap my brain around this. AD has nothing to do with intra-protocol route selection. Hope this helps, Bob On Tue, May 1, 2012 at 5:11 PM, Fulvio allegretti <[email protected]> wrote: Hi all, Doing lab 9 Volume 2 - Task 4.4, configure R6 such that if it learned the same route via eBGP, it would still prefer the same route it has from R2 (iBGP). For my little brain this was very clearly to do with distance, eBGP 20 - iBGP 200, change the distance for the eBGP peers the task refers to and job done. The solution suggests adjusting weight which has left me a bit confused, what do you think? I though that admin distance would be looked at before the path selection algorithm Using the same logic, why does step 7 of the algorithm uses eBGP and iBGP? Again, adming distance is different, so we shouldn't even start the path selection alogrithm. Unless the distance was changed manually I suppose. 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
