I support adoption of this draft as it provides an innovative feature of
 providing a new “soft data plane” for flex algo use cases independent of
existing SR Algo or IP Algo, Flex Algo data planes.

It allows for new flex algo link affinity bit position based sub topology
Algo participation in FADs with min delay or min bandwidth real time metric
constraints for any applications requiring low latency such as multicast or
minimum bandwidth  for app such as a  high bandwidth dc workloads.

Other examples could be NVO VXLAN, GENEVE DC, BGP Fabrics or even hybrid
cloud workflows requiring high bandwidth.

As Mankamana mentioned this draft extends multicast to now use Flex Algo
for multicast similar to mLDP MT aware path.

https://datatracker.ietf.org/doc/rfc9658/


Thanks

Gyan


On Mon, Oct 20, 2025 at 8:51 AM Peter Psenak <ppsenak=
[email protected]> wrote:

> Hi Acee,
>
> On 20/10/2025 14:34, Acee Lindem wrote:
> > Hi Peter,
> >
> >> On Oct 20, 2025, at 8:20 AM, Peter Psenak <[email protected]> wrote:
> >>
> >> Hi Acee,
> >>
> >> On 20/10/2025 13:58, Acee Lindem wrote:
> >>> Hi Les,
> >>>
> >>>> On Oct 17, 2025, at 11:29 AM, Les Ginsberg (ginsberg) <
> [email protected]> wrote:
> >>>>
> >>>> Esteemed WG chairs -
> >>>>
> >>>> What is the conclusion of the WG adoption call for
> draft-ginsberg-lsr-flex-soft-dataplane ??
> >>>>
> >>>> We are long past the intended deadline.
> >>>>
> >>>> I note that multiple people expressed support for adoption.
> >>> And a couple were even authors 😎
> >>>
> >>>> One person expressed concern about the document - which we tried to
> address when we published V2 of the draft on August 28, 2025.
> >>> That was me and I still don't see why the use case couldn't be
> satisfied with a constrained flex-algo in the multicast IP topology.
> >> Flex-algo is tight with the data-plane that is used to forward the
> traffic over the FA paths - e.g., to be able to compute FA paths, routers
> need to participate in one of the supported data-planes   - SR-MPLS, SRv6
> or IP flex-algo.
> > Well, in practice, implementations are not that strict about requiring a
> data plane for base IP computation and are using the mulitcast topology in
> a "soft" manner.
>
> flex-algo requires the participation in some DP, and that's why the
> "soft" one has been introduced.
>
> We did not want to restrict to the mcast use case. Mcast is just one
> app, there may be more apps in the future that may have the similar
> requirement for flex-algo.
>
> thanks,
> Peter
>
> >
> > Thanks,
> > Acee
> >
> >> The requirement here is to detach the flex-algo from any specific
> data-plane - application just needs the FA paths on top of regular IP
> topology and it will setup its own data-plane for forwarding over them.
> >
> >
> >> thanks,
> >> Peter
> >>
> >>
> >>>
> >>>> Please let us know next steps.
> >>> Nevertheless, we'll move forward with LSR WG adoption.
> >>>
> >>> Thanks,
> >>> Acee
> >>>
> >>>
> >>>> Thanx.
> >>>>
> >>>>    Les (on behalf of the authors)
> >>>>
> >>>>
> >>>>> -----Original Message-----
> >>>>> From: Acee Lindem <[email protected]>
> >>>>> Sent: Friday, August 22, 2025 2:23 PM
> >>>>> To: lsr <[email protected]>
> >>>>> Cc: Les Ginsberg (ginsberg) <[email protected]>; Peter Psenak
> (ppsenak)
> >>>>> <[email protected]>; <[email protected]>
> >>>>> <[email protected]>
> >>>>> Subject: LSR WG Adoption Poll for "IGP Flex Soft Dataplane" -
> draft-ginsberg-
> >>>>> lsr-flex-soft-dataplane-01
> >>>>>
> >>>>> LSR WG,
> >>>>>
> >>>>> The begins the LSR WG adoption call for "IGP Flex Soft Dataplane" -
> draft-
> >>>>> ginsberg-lsr-flex-soft-dataplane-01. Please express your support or
> objection
> >>>>> on this list prior to Saturday September 6, 2025.
> >>>>>
> >>>>> Thanks,
> >>>>> Acee
> >>
>
> _______________________________________________
> Lsr mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
>
_______________________________________________
Lsr mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to