All things related to multicast

[Toerless Eckert <>]

On Tue, Jun 12, 2001 at 04:11:49PM -0500, Bill Nickless wrote:
> It seems to me that there are five levels of efficiency for a multilateral 
> multicast exchange point, listed in order of (my) preference:
> 
> EFFECTIVELY SHARED MEDIA
> ========================
> Any participant gets all multicast traffic at the exchange point, whether 
> or not they want to get multicast traffic.  An example of this is an 
> exchange point built on FDDI or a multicast-ignorant Ethernet switch.
> 
> (*,*,*) SWITCHING
> =================
> Any participant who has PIM-SM configured gets all the multicast traffic at 
> the exchange point.  An example of this is the use of an IGMPv2 snooping 
> Ethernet switch.  I believe the NASA AMES MIX is at this efficiency level 
> based on Hugh LaMaster's earlier note in this thread.

Currently, we had RGMP running already, don't know why it's currently
disabled. What's the difference between (*,*,*) switching as you call
it and effectively shared media ? If you only have routers, and that's
what we're talking about in a MIX< it is effectively a shared media. You
won't be able to squeez more bits out of it except for a little bit
better buffering and avoidance of collision detection.

> (*,G,*) SWITCHING
> =================
> Participants only get multicast traffic for groups they are interested in 
> receiving.  Examples of this include RGMP (Cisco world) and the current 
> implementation of Foundry PIM-SM snooping.
> 
> (S,G,*) SWITCHING
> =================
> Participants only get multicast traffic for specific sources in specific 
> groups they are interested in receiving.  The Foundry PIM-SM snooping 
> feature may be extended to this level.

Then you could imagine a (*,G)+(S,G) switching...

> (S,G,RPF) SWITCHING
> ===================
> Participants get multicast traffic for specific sources in specific groups 
> from a single RPF peer (eliminating the possibility of packet duplication 
> if more than one participant is sourcing traffic to the exchange 
> point).  Three sub-cases:
> 
>    - Point-to-point bilateral PIM-SM peerings (using MPLS, ATM VCs,
>      VLANs, etc) meet this goal, but each participant must do their
>      own outbound packet replication.
> 
>    - The Toerless Eckert VLAN-per-participant scheme is an example
>      of how to implement this at the lesser cost of a single VLAN per
>      participant, while letting the exchange point do the packet
>      replication.  The cost is configuration complexity and sparse
>      use of exchange IP address space.
> 
>    - Having each participant peer with an exchange-point router over
>      a private VLAN also meets this goal but at the cost of delegating
>      the choice of RPF for any given source to the exchange point
>      managers.
> 
> Is this a reasonable taxonomy?

I think that what you call (S,G,RPF) switching is not for what
you say: elimination of duplication (that already should work in the
other solutions), but instead it is about the ability of each downstream
peer to individually slect it's upstream peer for a certain traffic
source. This of course doesn't apply to the lat case you cite, the
L3-MIX. In addition, it is i think orthogonal to the mechanisms described
above, eg: in the solution i proposed, you could use RGMP or PIM
Snooping in each VLAN too.

Finally, if you want to make a taxonomy, then it might be interesing
(although most likely not of much practial use), that the most
efficient and flexible method to set up a MIX is to use RFC2337, eg:
mapping of PIM to ATM Multipoint signalling. You use an ATM-switch
with ATM-Multipoint signalling and replication and each peer who
sources a (*,G) or (S,G) establishes an ATM point-to-multipoint SVC
to those peers on the mix interested in that stream from exactly this
upstream neigbhor. So you have the full multicast eficiency and flexibility:
Each upstream peer only needs to send out data once, and each downstream
peer only receives traffic from that upstream neighbor that it selected by
it's local routing decisions. What i proposed with ethernet and VLANs
achieves the same advantages but is of course more ugly to set up
- that's the price to pay for wanting to use inexpensive ethernet ;-))

Cheers
        Toerless