Source Specific Multicast (SSM)

This topic is to discuss the following lesson:

Dear Sir,
You explain in a very good way, your explanation make the tough things simpler. Thanks for your contribution to the Networking Society.

Thats a good explanation Rene.
But my concern is what about the need to enable ssm map.
When to use this map n all…

Hi Rene,
I am a beginner and interested in learning IGMP and PIM protocols . Can you please point me the links to start these protocols?

Thanks,
Srini

Rene,
Can we enable SSM with Dense-mode?

Regards,
Ulrich

Hello DJAN

When configuring SSM, you can use either sparse mode or sparse-dense mode. More information on configuring SSM can be found here:


I hope this has been helpful!

Laz

1 Like

Laz,
Thank for r answer.

Cheers!

Does PIM-SSM “flood” multicast between routers, or is the routing table built dynamically without flooding?

Using the word “flood” very loosely, we are trying to avoid having to forward all of our video streams to an RP, and need avoid any type of Dense mode or similar behavior; we are working with very high bit rate streams on 10G interfaces, up to 9Gbps for slightly compressed 4K60 4:4:4 Video.

Hi Vanya,

It’s not flooded. With sparse mode, we initially use an RP as a central point since we don’t know where the source is. Once we learn the source, we switch to SPT (unless you disable this).

With SSM, you have to specify the source so that’s why you don’t need the RP. Traffic is forwarded with sparse mode.

Rene

Thanks Rene!

Generally speaking, how long does the transition from RPT to SPT take? Is the timing configurable?

Regards,
Vanya

Hi Vanya,

SPT switchover is fast. It happens right after the first data packet is received so it’s not based on a timer.

The advantage of using source trees is that the delay is lower. After all, it is the shortest path from the source to a receiver. The disadvantage is that it requires more memory since you need to keep track of multiple states.

You can use the ip pim spt-threshold command to define when you want to switch to SPT. By default it’s set to 0 kbps so it switches right away after receiving the first data packet. You can also configure it never to switch to SPT or once you exceed a certain threshold (in kbps).

Rene

SSM need PIM neighbor to establish ?

Hello Devaprem

In order to enable SSM, PIM must be enabled. You must also be running IGMPv3 in order to enable SSM. You can find out more about it at the following lesson:

I hope this has been helpful!

Laz

Hi Laz,

This pdf link is no longer valid. Do you have a copy of the document elsewhere or can you summarize the purpose of the pdf?

Thank you,

Jay Karson

Hello Jay

Thanks for pointing that out, here’s the new link:

I’ve updated the above post as well.

Thanks again!

Laz

If I set up multicast as normal with an RP and dont use this command ip pim ssm default and configure the
multicast destination addresses in the 232 /8 range what will happen ?

Will the router know that I want to use SSM just by me using that 232 /8 address range or do you need the ip pim ssm default
command ?
I know SSM doesn’t use the RP obviously but I’m wondering what would happen ?

Is it mandatory to use igmp version 3 on the source port and destination port as Rene mentions
in the lesson if I am implementing this on L3 switches and also would this then need to be also added to all SVI’s
on those layer 3 switches.

If the destination was a server for example would this server normally be where the configuration of where he wants to source
the multicast traffic from be configured.

We have a large production network but I’m not sure multicasting is set up correctly.
All SVI’s are running pim sparse-dense mode
All are L3 switches configured with Auto RP and static config pointing to the RP so that is fine but
when I look at show ip mroute on any of the switches I can see all addresses are in the 232 /8 range so I then wonder the point of having the RP
but even more strangely the show IP mroute table shows all these 232 /8 are shown in a star,G state ie for example you may see something like
(*, 232.22.22.23) etc, and none show as S,G (source, group)

I may be wrong but even when the ssm traffic is traversing between SVI’s I would expect to see s,g state and not *,g

Then if I check something like show ip igmp snooping groups I can still see these addresses but the IGMP version shows as V1 and V2 which
further confuses me as I though you need igmp V3 for SSM.

Maybe its possible to use the 232/8 range for non ssm if we dont use the command ip pim ssm default ? but it would be strange for someone
to choose that range in the first place I think or can that range be used as “normal” multicast also ?
Also the outgoing interface list (oil) is empty on all svi’s on all switches.
they will all show multicast addresses in the (star,232/8) range and also empty oil.
If this multicasting is not working correctly is it possible the receiver is still getting the multicast through other means such as flooding etc. ?

Hello Sean

If you use the command ip ssm default, as in the lesson, then the 232.0.0.0/8 range is used by default. Alternatively, you can use the ip ssm range ACL_NAME where the the ACL_NAME is the name of the access list that defines the multicast address range you want to define. This way, you can use a range different from 232.0.0.0/8. In both cases, if you try to join a group outside of the specified range (either default, or otherwise specified) it will be dropped.

No, the router will not know you want to use SSM simply the the range being used. You must use the ip pim ssm command.

IGMPv3 is necessary for SSM to function, and all participating routed ports must have IGMPv3 enabled.

Yes, it is possible to use this range just like any other multicast group if you do not enable SSM. You can find some more info about SSM at this Cisco documentation:

I hope this has been helpful!

Laz

1 Like