[nsp] BGP, best way to build an transit AS

[Lukas Krattiger]

Thanks for all the usefull answers, I like to write/read this list ... 
always great ideas and answers :-)

And an other question/topic ...
         best way to build an transit AS

We are starting now to offer transit through our network, at this time we 
are doing bgp to some upstreams (4) and peering. We configured our prefixes 
with the network command (and aggregation), filter-list, prefix-list etc, 
see config-part below.

router bgp AS_ME
  no synchronization
  no bgp fast-external-fallover
  bgp log-neighbor-changes
  network X.X.X.X mask Y.Y.Y.Y
  aggregate-address X.X.X.X Y.Y.Y.Y summary-only
  neighbor upstream1 remote-as AS2
  neighbor upstream1 prefix-list A out
  neighbor upstream1 filter-list 10 out
  neighbor upstream2 remote-as AS2
  neighbor upstream2 prefix-list A out
  neighbor upstream2 filter-list 10 out
  neighbor customer1 remote-as AS_CUSTOMER
  neighbor customer1 filter-list 100 in
!
ip as-path access-list 10 permit ^AS_ME$
ip as-path access-list 10 permit ^AS_CUSTOMER$
ip as-path access-list 10 permit ^$
ip as-path access-list 10 deny .*
!
ip as-path access-list 100 permit ^AS_CUSTOMER$
!
ip prefix-list A seq 5 permit X.X.X.X/Y le 32
ip prefix-list A seq 10 permit customer_prefix/24 le 32

Now we have some customers who want to get internet over bgp from us ...
what would be the easiest or most used way to reach to goal to announce the 
customer_prefix and AS to my two upstreams. I have to add more transit 
customers in  soon so I want to find a generic way to add them "easy". :-)

Thanks in advance

-Lukas


At 16:39 04.08.2003, Lucas Iglesias wrote:
>You can do this with policy-routing (be careful, it can load your CPU).
>An example would be this one:
>
>route-map RM_OUTGOING1 permit 10
>  match ip address 1
>  set ip next-hop 192.168.1.1 192.168.1.5
>!
>route-map RM_OUTGOING2 permit 10
>  match ip address 2
>  set ip next-hop 192.168.1.5 192.168.1.1
>!
>
>Assuming that 192.168.1.1 is the next-hop from a + b and 192.168.1.5 is from
>c + d (the set ip next-hop allow you to put next-hops in order, if the first
>is not reachable use the next one).
>ACL 1 must permit one group of IP source nets, and ACL 2 the other one.
>Finally, you must apply the route-maps to all the incoming customer
>interfaces, i.e.:
>
>interface Serial5/0/1
>  ...
>  ip policy route-map RM_OUTGOING1
>end
>
>interface Serial5/0/2
>  ...
>  ip policy route-map RM_OUTGOING2
>end
>
>Good Luck.
>========================
>Eng. Lucas Iglesias
>IP Engineering, Tiba S.A.
>========================
>
>
>
>
>-----Mensaje original-----
>De: Lukas Krattiger [mailto:luk at everyware.ch]
>Enviado el: Lunes, 04 de Agosto de 2003 06:44 a.m.
>Para: cisco-nsp at puck.nether.net
>Asunto: [nsp] BGP, best way to balance outgoging traffic
>
>
>Greetings,
>
>We are currently running our BGP with 4 upstreams and anounceing 6 prefixes.
>The advertisement of this 4 prefixes over upstream a + b and 2 prefixes
>over upstream c + d are working fine.
>What's now the best way to tell the outgoing traffic from all of this 2
>prefixes must take upstream c + d and the other 4 prefixes are only allowed
>to use upstream a + b ?
>Any examples ?
>
>Best Regards
>-Lukas
>
>_______________________________________________
>cisco-nsp mailing list  cisco-nsp at puck.nether.net
>http://puck.nether.net/mailman/listinfo/cisco-nsp
>archive at http://puck.nether.net/pipermail/cisco-nsp/