mcproxy igmp V2 fast leave

130 views
Skip to first unread message

Ke Hu

unread,
Aug 9, 2019, 2:59:15 AM8/9/19
to multica...@googlegroups.com
Hey guys,

We are using Mcproxy in our openwrt router. It is so fancy and stable. But we can not find any feature like 'fast leave'. Is it not supported by Mcproxy? or Is there any configure that we can set?

BR

Thomas C. Schmidt

unread,
Aug 9, 2019, 7:16:42 AM8/9/19
to multica...@googlegroups.com
Hi Ke Hu,

great to hear!

I'm not sure what you mean by "fast leave". If you refer to the
IGMPv2/MDLv1 leave signaling (i.e., groups are actively left by clients
as opposed to timeouts in IGMPv1), this is the standard operation of
MCProxy. We do not support an IGMPv1-Mode.

Otherwise I'm not sure what you mean: for the proxy function, RFC 4605 says:
"The proxy device sends IGMP/MLD membership reports on the upstream
interface when queried and sends unsolicited reports or leaves when
the database changes."

This is immediate (so fast) and this is what MCProxy should do (unless
you find a bug).

Best,
Thomas
> --
> You received this message because you are subscribed to the Google
> Groups "Multicast Proxy" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to multicast-pro...@googlegroups.com
> <mailto:multicast-pro...@googlegroups.com>.
> To view this discussion on the web, visit
> https://groups.google.com/d/msgid/multicast-proxy/AM0PR08MB297776BFE35A5D0CFDF921B0F6D60%40AM0PR08MB2977.eurprd08.prod.outlook.com
> <https://groups.google.com/d/msgid/multicast-proxy/AM0PR08MB297776BFE35A5D0CFDF921B0F6D60%40AM0PR08MB2977.eurprd08.prod.outlook.com?utm_medium=email&utm_source=footer>.

--

Prof. Dr. Thomas C. Schmidt
° Hamburg University of Applied Sciences Berliner Tor 7 °
° Dept. Informatik, Internet Technologies Group 20099 Hamburg, Germany °
° http://inet.haw-hamburg.de/members/schmidt Fon: +49-40-42875-8452 °

kehu

unread,
Aug 9, 2019, 7:33:32 AM8/9/19
to Multicast Proxy
thanks for your kindly reply.

the 'fast leave' ,we would like to delopy in IGMPV2,  is when the last member in the group sending leave message. the router just sends leave message to upstream server without sending any specific 
query to downstream ports and  delete the multicast group in the router immediately. Although it is supported by our hardware chipset, we would like to use software method.
And if it is not supported by Mcproxy, it is possible to develop a patch to the current source code.

Thanks so much
------------------------------------
Hi Ke Hu,

great to hear!

I'm not sure what you mean by "fast leave". If you refer to the
IGMPv2/MDLv1 leave signaling (i.e., groups are actively left by clients
as opposed to timeouts in IGMPv1), this is the standard operation of
MCProxy. We do not support an IGMPv1-Mode.

Otherwise I'm not sure what you mean: for the proxy function, RFC 4605 says:
  "The proxy device sends IGMP/MLD membership reports on the upstream
    interface when queried and sends unsolicited reports or leaves when
    the database changes."

This is immediate (so fast) and this is what MCProxy should do (unless
you find a bug).

Best,
  Thomas

On 09/08/2019 08:56, Ke Hu wrote:
> Hey guys,
>
> We are using Mcproxy in our openwrt router. It is so fancy and stable.
> But we can not find any feature like 'fast leave'. Is it not supported
> by Mcproxy? or Is there any configure that we can set?
>
> BR
>
> --
> You received this message because you are subscribed to the Google
> Groups "Multicast Proxy" group.
> To unsubscribe from this group and stop receiving emails from it, send

Thomas C. Schmidt

unread,
Aug 10, 2019, 3:07:23 AM8/10/19
to multica...@googlegroups.com
Hi Ke Hu,

usually, the router does not know whether a leaving node is the last one
of this group. It can know, if it performs explicit tracking of
listeners, but if I remember correctly, this does not work with IGMPv2,
since IGMPv2 listeners suppress join messages if a group is already
requested by other nodes.

Do you have a pointer to the specification (RFC section) of what you
want to implement?

Best,
Thomas

On 09/08/2019 13:33, kehu wrote:
> thanks for your kindly reply.
>
> the 'fast leave' ,we would like to delopy in IGMPV2,  is when the last
> member in the group sending leave message. the router just sends leave
> message to upstream server without sending any specific
> query to downstream ports and  delete the multicast group in the router
> immediately. Although it is supported by our hardware chipset, we would
> like to use software method.
> And if it is not supported by Mcproxy, it is possible to develop a patch
> to the current source code.
>
> Thanks so much
> ------------------------------------
>
> > an email to multica...@googlegroups.com <javascript:>
> > <mailto:multicast-pro...@googlegroups.com <javascript:>>.
> <https://groups.google.com/d/msgid/multicast-proxy/AM0PR08MB297776BFE35A5D0CFDF921B0F6D60%40AM0PR08MB2977.eurprd08.prod.outlook.com?utm_medium=email&utm_source=footer
> <https://groups.google.com/d/msgid/multicast-proxy/AM0PR08MB297776BFE35A5D0CFDF921B0F6D60%40AM0PR08MB2977.eurprd08.prod.outlook.com?utm_medium=email&utm_source=footer>>.
>
>
> --
>
> Prof. Dr. Thomas C. Schmidt
> ° Hamburg University of Applied Sciences                  Berliner
> Tor 7 °
> ° Dept. Informatik, Internet Technologies Group   20099 Hamburg,
> Germany °
> ° http://inet.haw-hamburg.de/members/schmidt
> <http://inet.haw-hamburg.de/members/schmidt>      Fon:
> +49-40-42875-8452 °
>
> --
> You received this message because you are subscribed to the Google
> Groups "Multicast Proxy" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to multicast-pro...@googlegroups.com
> <mailto:multicast-pro...@googlegroups.com>.
> To view this discussion on the web, visit
> https://groups.google.com/d/msgid/multicast-proxy/dd323103-2533-474e-b20c-6ecee9a619f9%40googlegroups.com
> <https://groups.google.com/d/msgid/multicast-proxy/dd323103-2533-474e-b20c-6ecee9a619f9%40googlegroups.com?utm_medium=email&utm_source=footer>.

hk484...@gmail.com

unread,
Aug 10, 2019, 4:01:20 AM8/10/19
to multica...@googlegroups.com

thanks for the tips. The downstream is a bridge, in our case. So it becomes more difficult to know the last member. It probably needs to work with bridge snooping.

Concerning RFC, I can not find any description about this. Because we used to use broadcom platform. It did some trick to the kernel the proxy daemon, not standard, hacking way.
> To unsubscribe from this group and stop receiving emails from it, send an email to multicast-pro...@googlegroups.com.
> To view this discussion on the web, visit https://groups.google.com/d/msgid/multicast-proxy/0c214424-01e6-301a-e94e-ea94e579d3b2%40haw-hamburg.de.

Thomas C. Schmidt

unread,
Aug 10, 2019, 12:13:26 PM8/10/19
to multica...@googlegroups.com
Hi Ke Hu,

if a downstream bridge performs IGMP snooping, the upstream router would
still not know about the mcast table in the bridge. So this would not help.

In any case, McProxy is bound to implementing the standard proxy
behavior. We are not providing specific 'hacks' for tricky hardware.

Best,
Thomas

On 10/08/2019 10:01, hk484...@gmail.com wrote:
>
> thanks for the tips. The downstream is a bridge, in our case. So it becomes more difficult to know the last member. It probably needs to work with bridge snooping.
>
> Concerning RFC, I can not find any description about this. Because we used to use broadcom platform. It did some trick to the kernel the proxy daemon, not standard, hacking way.
>
>> 在 2019年8月10日,09:07,Thomas C. Schmidt <t.sc...@haw-hamburg.de> 写道:
>>
>> To unsubscribe from this group and stop receiving emails from it, send an email to multicast-pro...@googlegroups.com.
>> To view this discussion on the web, visit https://groups.google.com/d/msgid/multicast-proxy/0c214424-01e6-301a-e94e-ea94e579d3b2%40haw-hamburg.de.
Reply all
Reply to author
Forward
0 new messages