Robert,

That's why I replied it depends, I wanted to ensure Zaidoon was aware of
scenarios
where it was appropriate.

Mike

On Tue, Aug 23, 2011 at 9:02 AM, Robert Raszuk <[email protected]> wrote:

> Well of course.
>
> But I assumed that the question is about connecting the given pair of
> ASBRs over parallel physical links between them.
>
> R,
>
> > Depends, if the two ISP peers were located at two different POPs and your
> > layer one connectivity
> > was diverse this would help your AS in more failure scenarios than a
> single
> > threaded design. Of course
> > I would also diversify the connections onto different linecards/slots as
> > well.
> >
> > Mike
> >
> >
> >
> > On Tue, Aug 23, 2011 at 7:58 AM, Robert Raszuk <[email protected]>
> wrote:
> >
> >> Hi zaidoon,
> >>
> >> Nope - I would not recommend that.
> >>
> >> Your better choice is to peer between loopbacks and use
> >> disable-connected-check knob or BGP multihop.
> >>
> >> Two sessions will cause you to get the same paths two times wasting a
> >> bit of control plane memory and CPU inbound processing - but that's
> >> about it. On the peer's side update generation would be the same as your
> >> peer would only copy at replication. But better is to have single
> >> session IMHO.
> >>
> >> Cheers,
> >> R.
> >>
> >>> Is it  recommended to terminate
> >>>  two bgp session on 12000 xr that  peering with the same isp on the
> same
> >> router how
> >>> to handle full routing table ? Any clues
> >>>
> >>> Zaid
>
>
_______________________________________________
cisco-nsp mailing list  [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

Reply via email to