RE: BGP next-hop

Hi Shane,
 

I know that there is three possible solutions to influence the next-hop problem in BGP
First, we could advertise the next-hop through our IGP/redistribute it
Second, we can use as next-hop-self, as long as we have route to what we are sourcing the update from (update-source command) we are ok
Third, we can change the next-hop through route-map, and again as long as the router can make full recursive to this next-hop we are ok!

   EBGP IBGP
R1-------R2--------------R3
    S
 

My question, if we have the following scenario, R2 advertise the S subnets through BGP network statment
or redistribute the connected into BGP, R3 can only know S subnet through BGP itself not IGP.. then
R3 can make full recursive lookup and we are ok !
I Tested it, It worked fine,and thats why I'am asking because i saw alot of the documents mentioned that
you should know it through IGP only !
If we know S subnet through BGP & IGP, I know we gonna choose the IGP based on the AD.
 
  Thanks
 Subject: Re: BGP next-hop
From: shane_at_shanekillian.net
Date: Thu, 26 Jul 2012 21:12:43 +0100
CC: mohd-mousa_at_hotmail.com; ccielab_at_groupstudy.com
To: rtvrpk_at_gmail.com

For what its worth, just happened to be lambing this when the mail came in.The answer is no, you must use the next-hop-self command
Rack1R4(config-router)#neighbor 155.1.0.1 next-hop-self

Without next hop self -

Rack1R1#show ip bgp 112.0.0.0 255.0.0.0BGP routing table entry for 112.0.0.0/8, version 24Paths: (2 available, best #2, table default) Not advertised to any peer 54 50 60 54.1.1.254 (metric 2195456) from 155.1.146.6 (150.1.6.6) Origin IGP, metric 0, localpref 100, valid, internal 54 50 60 204.12.1.254 (metric 307200) from 155.1.0.4 (150.1.4.4) Origin IGP, metric 0, localpref 100, valid, internal, bestRack1R1#Rack1R1#Rack1R1#Rack1R1#show ip route 112.0.0.0Routing entry for 112.0.0.0/8 Known via "bgp 100", distance 200, metric 0 Tag 54, type internal Last update from 204.12.1.254 00:03:42 ago Routing Descriptor Blocks: * 204.12.1.254, from 155.1.0.4, 00:03:42 ago Route metric is 0, traffic share count is 1 AS Hops 3 Route tag 54 MPLS label: noneRack1R1#Rack1R1#Rack1R1##### Recursion to get to next hop advertised ####Rack1R1#show ip route 204.12.1.254Routing entry for 204.12.1.0/24 Known via "eigrp 100", distance 90, metric 307200, t!
 ype internal Redistributing via eigrp 100 Last update from 155.1.146.4 on Ethernet0/0, 00:04:49 ago Routing Descriptor Blocks: * 155.1.146.4, from 155.1.146.4, 00:04:49 ago, via Ethernet0/0 Route metric is 307200, traffic share count is 1 Total delay is 2000 microseconds, minimum bandwidth is 10000 Kbit Reliability 255/255, minimum MTU 1500 bytes Loading 1/255, Hops 1Rack1R1#

With Next hop self -

Rack1R1#show ip bgp 112.0.0.0 255.0.0.0BGP routing table entry for 112.0.0.0/8, version 32Paths: (2 available, best #2, table default) Not advertised to any peer 54 50 60 54.1.1.254 (metric 2195456) from 155.1.146.6 (150.1.6.6) Origin IGP, metric 0, localpref 100, valid, internal 54 50 60 155.1.0.4 from 155.1.0.4 (150.1.4.4) Origin IGP, metric 0, localpref 100, valid, internal, bestRack1R1#
Rack1R1#show ip route 112.0.0.0Routing entry for 112.0.0.0/8 Known via "bgp 100", distance 200, metric 0 Tag 54, type internal Last update from 155.1.0.4 00:00:32 ago Routing Descriptor Blocks: * 155.1.0.4, from 155.1.0.4, 00:00:32 ago Route metric is 0, traffic share count is 1 AS Hops 3 Route tag 54 MPLS label: noneRack1R1#

On 26 Jul 2012, at 19:24, Ray wrote:When using iBGP, it is assumed that there will be an underlying IGP to
control/forward advetsisments. iBGP has AD of 200 and will typically lose
out ot other routing protocols by design. It was set up this way as a loop
prevention mechanisim. Like stated above, a route learned from outside the
Internetwork is assumed to be valid best. This is not the case when the
same route is learned internal to the AS.

So the above is default behavior. We can peer iBGP neighbors with the
'next-hop' self command or use a route-map attached to the neighborship
that changes next hop. It depends on what is trying to be accomplished.

/ray

On Thu, Jul 26, 2012 at 5:10 PM, <mohd-mousa_at_hotmail.com> wrote:

Hi All,

There is question for bgp next-hop, can we learn the next-hop through the
bgp
itself not through IGP ?

Could anyone provide me with example why we can't do this, I tried to find
example but no luck !

Thanks Alot,

Blogs and organic groups at http://www.ccie.net
Received on Fri Jul 27 2012 - 17:10:28 ART