[c-nsp] BGP transit selection from source customer network

[Alexandre Durand]

weird I used only PBR and it seems to be working. I created a route-map 
that send cust1 traffic out to ISPA (RIB internet routes) and any other 
customer source traffic out to ISPB. the weird thing is that ISPB routes 
are NLRI routes and not RIB routes ... how can this be working ...

regards

alexandre

route-map cust1, permit, sequence 10
   Match clauses:
     ip address (access-lists): cust1
   Set clauses:
     ip next-hop ISPA
   Policy routing matches: 122 packets, 13908 bytes
route-map cust1, permit, sequence 20
   Match clauses:  (ANY)
   Set clauses:
     ip next-hop ISPB
   Policy routing matches: 15 packets, 1710 bytes

On 02/08/11 11:04, Alexandre Durand wrote:
> Hi,
>
> You are right, however on router A is connected to ISPA, bgp will add 
> only preferred ISPA routes and  not the others routes from router B 
> ISPB with IBGP, but only ISPA routes will be preferred on this router 
> A (attribute weight), router A will be not aware about ISPB rouites 
> unless ISPA goes down.
>
> I may use a dedicated vrf for ISPA so global routing table and vrf 
> table are not bound each other. And then use PBR over route-map to 
> redirect first via ISPA VRF and secondly via global routing table 
> where ISPB routes will be present in RIB.
> What do you think about it?
>
> The only issue witch such solution is that I will need to upgrade the 
> ios because vrf under route-map is not supported on my IOS(12.2(18)SXD3).
>
> regards
>
> Alexandre
>
> On 01/08/11 17:56, David Freedman wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 01/08/11 15:35, Alexandre Durand wrote:
>>> Hi Andrew,
>>>
>>> Tnhan you for you answer. Actually youa re right about the inbound
>>> traffic in which I can easily advetise this network with a 
>>> preference to
>>> my "ISPA", this is fine and already done.
>>>
>>> The issue here is about the oubound traffic, when my customer 
>>> traffic is
>>> going back out to ISP, I can also you weight on the router where I host
>>> this "ISPA" and use no-advertise attribute so I don t advertise
>>> prefixes to others iBGP routers, fine but others customer traffic will
>>> also be routed via ISPA because the shortest ISP path is going throught
>>> this router too...
>>>
>> I think what you are saying sounds like this:
>>
>> |Your Network---Internet---------|
>>
>> [Cust1]--( N )
>>           ( E )--[ISPA]-(N)-[Cust4]
>> [Cust2]--( T )         (E)
>>           ( W )--[ISPB]-(T)-[Cust5]
>> [Cust3]--( K )
>>
>>
>>
>> I take it you want to do :
>>
>> - - Predetermined outbound path Cust1->Cust4 via ISPA
>> - - Predetermined outbound path Cust2->Cust5 via ISPB
>> - - Cust3 ->  Cust4/5 via any/best path
>>
>> What you are talking about is essentially policy routing at your edge
>> (i.e using your own criteria, like customer ID to influence outbound
>> routing, not metrics)
>>
>> you can do this simply with PBR at the edge but it won't be very
>> reliable, but you might want to look at OER/PFR
>>
>> (http://www.cisco.com/en/US/products/ps8787/products_ios_protocol_option_home.html) 
>>
>>
>> Dave.
>> -----BEGIN PGP SIGNATURE-----
>> Version: GnuPG v1.4.10 (GNU/Linux)
>> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>>
>> iEYEARECAAYFAk42zL4ACgkQtFWeqpgEZrK2BgCgwGU7nRzRPklKNII9L69uA00I
>> w5YAoKkhzi8L8Roo5oONsg/6z1aNW0ka
>> =lLQt
>> -----END PGP SIGNATURE-----
>> _______________________________________________
>> 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/
>
>


-- 
Alexandre DURAND
TAS FRANCE
WTC 1-K, 1300 route des Crêtes
06560 Valbonne Sophia Antipolis
Phone :    +33 (0)4 92 94 56 93
Fax   :    +33 (0)4 92 94 33 99
Web:       http://www.tasfrance.com
Email :    alexandre.durand at tasfrance.com
peering:   http://as8554.peeringdb.com