BGP support

[AHMADJ]

please see  attached network  diagram

please note i have full mesh IBGP on same AS 47 between R1 AND R2 AND R3 AND R4,I HAVE one query to connect two customers to R3and R4  with different AS numbers and different location using EBGP

R3 WILL BE CONNECTED TO AS 1200 (EBGP)
R4 WILL BE CONNECTED TO AS 1000 (EBGP)

my question once EBGP is up the new customer prefixes added in R4 and R4 IT WILL BE LEARNED BY BOTH ROUTERS (r3 and r4 )AND VISE VERSA?if so what is the way to avoid it?and same will be learned in R1 and R2?

[MHM Cisco World · Accepted solution]

Hi

From my understand you ask about loop avoidance in ibgp, 

Friend, each peer will see same perfix but different next hop and it will use best path algor to select only one. 

[paul driver]

Hello
Bgp by default will only install in the rib the best path to any destination which will be based on the bgp path selection criteria – here

However you can manipulate this to your own preference
The first part of the example would to negate your AS becoming a transit path for either ISP, by just advertising locally originated routes to either isp and not any isp routes and also advertise itself as the next-hop to those egress routes
R3
ip as-path access-list 10 permit ^$
neighbor 1.1.1.1 next-hop-self
neighbor <isp as1200> filter-list 10 out

R4
ip as-path access-list 10 permit ^$
neighbor 2.2.2.2 next-hop-self
neighbor <isp as1000> filter-list 10 out

As I stated you can manipulate the prefer egress /ingress path with bgp attributes.

The example below would make r3 the preferred egress/ingress rtr for all ibgp peers, and advertised to the isps a probable ingress path into your AS.

r3 < egress path
router bgp 47
bgp default local-preference 200

r4 <ingress path from ISP
route-map pre-pend permit
set as-path prepend 47 47 47

router bgp 47
neighbor <isp as1000> route-map prepend out

The above is a very simple example for egress/ingress bgp path manipulation, There is much more that can be accomplished based on bgp path attributes.