[c-nsp] Changing from iBGP to eBGP - Again

[Oliver Boehmer (oboehmer)]

so you have two routers per Pop, each of which needs to peer with one
provider router in the same pop? So you would end up with 4 eBGP
connections in total? And you have four routers in your own ASN which
speak BGP? If this is the case, you will have one eBGP peer and three
iBGP peers on every router:

router bgp <your-new-asn>
 no synch
 neighbor <prov-router-1> remote-as ISP-ASN
 neighbor <prov-router-1> prefix-list myownspace out
 neighbor IBGP-MESH peer-group
 neighbor IBGP-MESH remote-as <your-new-asn>
 neighbor IBGP-MESH update-source loopback0
 neighbor IBGP-MESH next-hop-self
 neighbor a.a.a.a peer-group IBGP-MESH
 neighbor b.b.b.b peer-group IBGP-MESH
 neighbor c.c.c.c peer-group IBGP-MESH

ip prefix-list myownspace permit x.x.x.x/16 le 24

all the "network" statements which you might currently use to advertise
the address ranges stay in place..

But without configs, this is a only a somewhat educated guess..

	oli


Mark Tohill <mailto:Mark at u.tv> wrote on Wednesday, July 06, 2005 6:55
PM:

> Oliver,
> 
> Our current setup look something like:
> 
> 
> PoP No.1                                 PoP No.2
>                          WAN
>    |----------------------------------------|
> 
>    |                                        |
> PriRouter ---LAN----- SecRouter            PriRouter ---LAN---
>    SecRouter |         [HSRP]       |                  |         
>    [HSRP]    | |                      |                  |           
>   | ibgp                   ibgp               ibgp                
>    ibgp |                      |                  |                  
>    | |                      |                  |                    |
> ProviderRtr1---LAN----ProviderRtr2
> ProviderRtr1---LAN----ProviderRtr2
>                 |                                         |
>                 |                                         |
>            ProviderRtrX                               ProviderRtrX
>                 |                                         |
>                 | ----------------------------------------|
>                                    WAN
> 
>      Provider                                             Provider
>      Pop No.1                                             Pop No.2
> 
> 
> I hope that makes sense!
> 
> Following points:
> We announce chunk of our class B from each of our PoP's
> (non-contiguous) 
> 
> Neighbor relationships between us and provider are all iBGP
> 
> We are not implementing route-reflectors, communities, confederations
> etc.,
> just have multiple neighbor statements, alot of repetition. Probably
> about        six ibgp neighbors per router.
> 
> Synchronization not running
> 
> Will I post configs....?  :)
> 
> Thanks
> Mark
> 
> 
> 
> 
> 
> 
> 
> 
> 
> -----Original Message-----
> From: Oliver Boehmer (oboehmer) [mailto:oboehmer at cisco.com]
> Sent: 06 July 2005 16:50
> To: Mark Tohill; cisco-nsp at puck.nether.net
> Subject: RE: [c-nsp] Changing from iBGP to eBGP - Again
> 
> Mark,
> 
> it would be best if you pasted your current BGP config and topology so
> we know what you are doing today. I assume you have an iBGP peer to
> your upstream, and do route-reflector-client to your iBGP peers in
> order to reflect the upstream routes to them and vice versa.
> Changing this to an eBGP config should be easy, you wouldn't need
> RR-client configuration anymore, and you want to investigate
> next-hop-self on your iBGP peers (possibly not a requirement as you
> had to advertise the upstream's next-hop into your IGP previously).
> 
> Or do you also need to change your ASN, so you'd need to touch all
> routers? This is a larger migration task than just change one
> neighbor.. 
> 
> please clarify..
> 
> 	oli
> 
> 
> Mark Tohill <> wrote on Wednesday, July 06, 2005 5:37 PM:
> 
>> I realize this is probably against rules of list but I'm reposting
>> this during what seems a busy time.
>> 
>> 
>> 
>> Any help appreciated.
>> 
>> 
>> 
>> Thanks,
>> 
>> Mark
>> 
>> ________________________________
>> 
>> From: Mark Tohill
>> Sent: 05 July 2005 11:32
>> To: cisco-nsp at puck.nether.net
>> Subject: Changing from iBGP to eBGP
>> 
>> 
>> 
>> Hi,
>> 
>> 
>> 
>> We're hoping to change our setup with our sole up-stream provider
>> (peered with them across 2 x  PoPs) from iBGP to eBGP. We hope this
>> will give us more flexibility and control in the future
>> 
>> 
>> 
>> -          we accept the full routing table.
>> 
>> -          advertise non-contiguous (!!!) portions our our address
>> space from each of our PoP's. 
>> 
>> -          already secured our own ASN from RIPE.
>> 
>> -          have out-of-date route objects in RIPE which we are hoping
>> to amend. (or our provider, they maintain them)
>> 
>> -          hope to request this change fairly soon.
>> 
>> 
>> 
>> 
>> 
>> Has anyone out there hands-on experience of same/similar?
>> 
>> 
>> 
>> Apart from neighbor configuration (remote-as etc...), are there any
>> other major config changes? 
>> 
>> 
>> 
>> Are there any potential pitfalls?
>> 
>> 
>> 
>> Thanks,
>> 
>> Mark
>> 
>> _______________________________________________
>> cisco-nsp mailing list  cisco-nsp at puck.nether.net
>> https://puck.nether.net/mailman/listinfo/cisco-nsp
>> archive at http://puck.nether.net/pipermail/cisco-nsp/