Multicast immediate leave support in MS355

[jmcquillen]

Could someone please advise on support for Multicast immediate leave in MS355-24X2 running 11.22?

[PhilipDAth]

I wonder if @m_Andrew is about to help with this one.

[Kapil]

Hello!

While we don't specifically call it "immediate-leave", a Meraki switch, by default will immediately remove a host from the recipient list as soon as it hears an IGMP leave message from the host. In short, a multicast stream is immediately stopped from being forwarded to a port when an IGMP leave message is received on the port.

[lukewarms]

Hi @Kapil 

Is there any Meraki documentation that states that sates that a Meraki switch, by default, will immediately remove a host from the recipient list as soon as it hears an IGMP leave message? We are having issues with slow switching between Crestron NVX multicast streams. Creston say that it is because Meraki do not support fast-leave/immediate leave. I cannot find any official Meraki documentation to say that it does. Just the odd discussion statement like this one. 

Thanks

Luke

[bcdaus]

I have wiresharks gathered on MS350 switches running MS12.27 that show this is not true.  MCAST will continue from the initial address and cause the port to send out both streams until the first stream times out.

This is consistent with Fast-Leave / Immediate-Leave not being enabled or correctly implemented.

In a Crestron NVX system you can turn the stream bandwidth down from 950Mbps to 450Mbps to get around this limitation assuming you are happy with the reduced video quality.

[etb]

Hello,

Sorry to necro this old thread - I can start a new thread if desirable.  But I basically have the same question regarding whether Meraki MS switches necessarily support a fast/immediate leave.

In my particular case, I have a large 3rd party audio/video manufacturer (Extron) saying that they have tested their video-over-IP solution with Meraki switches as recently as this year, and they found Meraki switches to not work with their products due to a lack of support for fast/immediate leave.

(The reason fast/immediate leave is so critical for these video-over-IP solutions is that individual video streams can be solidly in excess of 500Mbps.  So it can exceed the bandwidth of a 1Gbps port if the first multicast stream is not stopped very quickly when switching to receive a different multicast stream.  bcdaus and lukewarms above are referring to a very similar system from Extron's competitor, Crestron.  And bcdaus states that their testing seemed to show that the Meraki switch is not successfully conducting fast/immediate leaving.)

I've tried googling various combinations of related terms, and all relevant results I find are this thread plus some reddit threads.  I've also looked through the Meraki documentation I can find (link) and (link), but I can't seem to find anything definitive about fast/immediate leave.

So is there some documentation which officially states that Meraki MS in fact support a mechanism to immediately stop a multicast stream when an IGMP leave message is received?

[GIdenJoe]

I haven't found anything in the documentation stating support for immediate leave.
This could be a good case to run Catalyst switches in Meraki monitoring mode instead of native MS management mode to have support for "advanced features".

If the documentation is not complete it makes sense to do some experiments and pcaps.
Catalyst switches do support full IGMP features.

[GIdenJoe]

I do remember from a few years back that I had a hotel with a TV system that was multicast.  The overlap where both streams were present on a port were minimal.  I  however don't remember if the switch sent out IGMP group specific queries when the tv sent a leave for that channel.

It's quite hard to troubleshoot this without having proper gear that can quickly switch between streams.

[etb]

Interesting to know.  I don't suspect that Kapil was wrong/lying, so I assume that there is some kind of mechanism to stop an unwanted stream after not too terribly long.  TV channels are going to be significantly lower bandwidth though, so it probably didn't matter nearly as much as these video-over-IP streams which can practically saturate a 1 Gbps link with 1 stream.

But when there are 2 different large video-over-IP manufacturers and also community members seeing issues, it would be an awfully unstable limb for me to go out on myself.  Extron actually advised against using their products with Meraki switches, whereas Crestron notes that they have some installations but that there can be some suboptimal experiences.  So I was hoping that maybe Meraki might say that Extron and Crestron just had suboptimal switch configurations, and actually MS switches are  in fact documented and tested as immediately stopping a stream when no longer wanted.

Either way, thanks for your input.