Re: [pim] [Mcast-wifi] Issues with MLD and Android powersaving / sleep mode...

[Hitoshi Asaeda <[e-mail]>]

Hi,

> On Mar 29, 2020, at 4:33, Linus Lüssing <[e-mail]> wrote:
> 
> On Fri, Mar 27, 2020 at 07:19:34PM +0900, Hitoshi Asaeda wrote:
>> RFC6636 (Tuning the Behavior of the Internet Group Management Protocol (IGMP) and Multicast Listener Discovery (MLD) for Routers in Mobile and Wireless Networks) gives some guideline to tune IGMP/MLD.
>> Android and iOS implement some specific mechanisms to save battery power, hence I don't know if they work what we expect, though.
> 
> Oh, thanks! I didn't know about this RFC before. It's an interesting
> read.
> 
> It's maybe not 100% what you were thinking of with "Explicit
> Tracking of Membership Status", but it makes me wonder if the
> following approach could help:
> 
> (1) We already track host and not just port based listener state
>    for the Linux bridge "multicast-to-unicast" conversion feature
> (2) We could avoid timing out host based multicast listener
>    entries. And only remove such an entry in the following two
>    cases:
>    (i): The host vanishes in the Linux bridge FDB or
>         even only if it vanishes on the 802.11 layer
>    (ii): If an MLDv2 Report was received, but did
>          not contain the specific multicast group anymore
>    (iii): An MLDv1 Done was received for the specific group

Interesting discussions.

It might be a different topic, but in fact, there is an explicit tracking draft (*) expired about a few years ago. It failed IESG review because they said the specific experiments were not explicitly described in the draft even though its intended status was experimental.
(*) https://tools.ietf.org/html/draft-ietf-pim-explicit-tracking-13
It'd be great to resume the activity if someone who has the experiments of the explicit tracking function could give some inputs for this function from implementors'/operators' points of views (or even better work together for this document).

Regards,

Hitoshi


> Regards, Linus