[prev in list] [next in list] [prev in thread] [next in thread] 

List:       juniper-nsp
Subject:    Re: [j-nsp] BGP load-balancing
From:       Harry Reynolds <harry () juniper ! net>
Date:       2009-03-25 22:08:37
Message-ID: FE7C265AC873F446A623E39A19A0BEE7BC1619 () p-embx58-sac ! jnpr ! net
[Download RAW message or body]

Coming in late, but I feel the types of question you ask indicate your
preparing well. ;)

That said, I think a FT LB policy is best here, and consider the
omission a piece of errata. If we were not advertising a single rip
summary than perhaps the default per-prefix would apply. As documented
we have two control plane next hops, but a single data plane one. It
would be best to have two at both levels.

Regards

 

-----Original Message-----
From: juniper-nsp-bounces@puck.nether.net
[mailto:juniper-nsp-bounces@puck.nether.net] On Behalf Of Aamir Saleem
Sent: Wednesday, March 25, 2009 1:02 AM
To: Patrik Olsson; Juniper Puck
Subject: Re: [j-nsp] BGP load-balancing

Dear all,

thanks for your inputs.

Regards.
Aamir

On Wed, Mar 25, 2009 at 1:58 AM, Patrik Olsson <df@webkom.se> wrote:

> Hello!
>
> bgp multipath only enables per prefix loadbalancing.
> To achieve per flow loadbalance aswell you need to apply a per packet 
> loadbalance policy under routing-options forwarding-table.
>
> Cheers
> Patrik
>
>
> Aamir Saleem wrote:
> > Hi All,
> >
> > I need your input regarding BGP loadbalancing with multipath option.
> >
> > In JNCIP-M study guide, one of the requirement given in iBGP case 
> > study
> is
> > to "Redistribute a summary of the RIP routes into IBGP from both r6 
> > and
> r7"
> > and in second requiement "r5 must IBGP load-balance to the summary 
> > route representing the RIP prefixes".
> >
> > After advetiseing RIP summary route (192.168.0.0/22) from R6 and R7.

> > R5 router is able to see two viable next-hops after enabling 
> > multipath in R5 ibgp group but in forwarding table of R5 shows only 
> > one path via R6 is available.
> >
> > My question is:
> >
> > Is only enabling multipath option will achive above mention 
> > requirement regarding BGP loadbalancing ?
> >
> > As junos perform per-prefix loadbalncing by default so the iBGP case
> study
> > requiremnt that "r5 must IBGP load-balance to the summary route
> representing
> > the RIP prefixes" cannot be achiedved by with multipath only. All 
> > other router behind R5 router also following the same path as chosen
by R5.
> >
> > some output from test senario are below:
> >
> > aamir@LAB1 <aamir@LAB1># run show route protocol bgp logical-router 
> > R5
> > 192.168.0.0/22
> > inet.0: 80 destinations, 136 routes (80 active, 0 holddown, 0 
> > hidden)
> > + = Active Route, - = Last Active, * = Both
> > 192.168.0.0/22     *[BGP/170] 01:50:49, localpref 100, from 10.0.9.6
> >                       AS path: I
> >                     > to 10.0.8.6 via lt-1/2/0.116
> >                       to 10.0.8.10 via lt-1/2/0.117
> >                     [BGP/170] 01:50:49, localpref 100, from 10.0.9.7
> >                       AS path: I
> >                     > to 10.0.8.10 via lt-1/2/0.117 [edit] 
> > aamir@LAB1#
> >
> > Forwarding table:
> > Logical router: R5
> > Routing table: inet
> > Internet:
> > Destination        Type RtRef Next hop           Type Index NhRef
Netif
> > 192.168.0.0/22     user     0                          indr  262148
> 3
> >                                          10.0.8.6             ucst
974
> > 7 lt-1/2/0.116
> >
> >
> > Any input is more then welcome.
> >
> > Regards.
> > Aamir
>  > _______________________________________________
> > juniper-nsp mailing list juniper-nsp@puck.nether.net 
> > https://puck.nether.net/mailman/listinfo/juniper-nsp
>
>
_______________________________________________
juniper-nsp mailing list juniper-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp
[prev in list] [next in list] [prev in thread] [next in thread]