Hi Jaeb,
> Thank you for the reply, as always.
>
> I was worried the ideal suggestion would be something along those lines.
> I'm looking at getting BGP traffic to traverse the dataplane and handle it
> via an application, such as Quagga or FRR, then pass it back down to VPP and
> out through the appropriate interface, without needing a code rewrite of
> either one. However, ideally in the future I'd be convenient if certain data
> could be handled by an application without needing to rewrite each one.
>
> Is there no other way to punt traffic to the Linux host, perhaps via some
> interface, and allow it to be sent back down again? -=-=-=-=-=-=-=-=-=-=-=-
1) You could do something like this:
https://github.com/FRRouting/frr/wiki/Alternate-forwarding-planes:-VPP
2) For BGP you could also run with a completely separate packet path, e.g.
using TAP interfaces. You’d still need some way of programming routes into VPP
of course.
3) I spoke to one of the FRR maintainers last week. They are doing a bit of
reshuffling if I remember correctly, but there are already integrations with
other stacks, so adapting FRR directly with VPP should be something we should
do regardless.
4) I was planning of adapting exaBGP to run directly over VCL. Haven’t gotten
around to it (yet).
Cheers,
Ole
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#12730): https://lists.fd.io/g/vpp-dev/message/12730
Mute This Topic: https://lists.fd.io/mt/30962708/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
