Re: [pim] PIM call for adoption of draft-eckert-pim-rfc1112bis

["Stevens, Jim A Collins" <[e-mail]>]

> -----Original Message-----
> Date: Thu, 6 Jul 2023 19:57:04 +0200
> From: Toerless Eckert <[e-mail]>
>
> Dear PIM WG, Chairs
>
> At IETF116 i presented this draft and think it received good WG interest, the
> chairs suggestested we might want to do an adoption call before IETF117. So
> hereby sending this email reminding of this draft.
>
> Quick summary why ?
>
> RFC1112bis IMHO would complete our IGMP/MLD update process, which
> othrerwise is
> hopefully progrssing nicely with RFC3810bis, RFC3376bis and RFC3228bis
> going into
> IETF/IESG review.
>
> What is RFC1112bis - summary:
>
> RFC1112 defines "IP Multicast (ASM)" service (for IPv4 only) - AND IGMPv1
>
> We need to retire IGMPv1:
>    RFC1112bis removes the IGMPv1 text.
>    We can then make RFC1112 and hence IGMPv1 historic and therefore retire
> IGMPv1
>    Because of RFC1112bis, this will not retire the ASM service model which is
>    then defined via RFC1112bis.
>
> We want to declare IP Multicast (ASM) service to also apply to IPv6
>    There is no other document currently doing all this.
>    So RFC1112bis has all the changes to talk about IP and IPv6.
>
> Thats pretty much the changes summary. Threre are details of course,
>
> Hope we can get this also done in short time in the WG.
>
> Thanks so much
>     Toerless
>

I apologize for not reading this draft sooner.  I think it is a good draft and should be moved forward.  However, the following sentence in section B.4 should be changed from

"In result, ASM is today only" a recommended solution for controlled networks including controlled federated networks for applications for which SSM is not usable."

to something like

"In result, ASM is today "only" a recommended solution for controlled networks including controlled federated networks for applications for which SSM is not usable or that support applications with many-to-many communication patterns can create more (S,G) state than is feasible for networks to manage."

Rationale, this is consistent with RFC 8815 Section 4.4 that describes cases where ASM is a better solution than SSM.

As an example, this includes large MANETs and networks of MANETs with many-to-many communications between hundreds to thousands of nodes that are dynamically joining and leaving.

I left in the phrase "controlled federated networks" but don't understand what it means.  Can you define this phrase or give an example?  and can there be controlled networks that aren't federated?  and federated networks that aren't controlled?  If yes, what is a controlled network and what is a federated network?

Also, the sentence does not mention intradomain versus interdomain - which is the focus of RFC 8815.   Depending upon the definition of "controlled federated networks", should that phrase be replaced by "intradomain"? If yes, then maybe the sentence should be replaced by

"In result, ASM is today "only" a recommended solution for intradomain including controlled federated networks for applications for which SSM is not usable or that support applications with many-to-many communication patterns can create more (S,G) state than is feasible for networks to manage."

Also, the title of section B.4 is "Standard for IP multicasting in controlled networks".  So depending upon what "controlled networks" means, either (a) the title ought to change to intradomain networks and replace controlled networks by intradomain networks or (b) add a similar section to B.4 for standard for IP multicasting in intradomain networks.

Jim Stevens