All things related to multicast

[John Zwiebel <>]

On May 22, 2004, at 9:18 PM, Ross Finlayson wrote:

> > We need to have a standard directory that lists multicast sources.
>
> No we don't.  Do we need "a standard directory that lists JPEG 
> images"?  Do we need "a standard directory that lists MPEG movies"?  
> Of course not.

For a "real" channel, where that channel is always up and running and
never moves, you are totally right.

> All we need is people advertising their SSM sessions on their own web 
> sites - as links to SDP files (i.e., using the MIME type 
> "application/sdp").  There are already plenty of ways to find things 
> on web sites (i.e., search engines).

However, folks claim multicast will be useful for other applications
than internet TV.  (I'd like to think so, but other than Access Grid,
it isn't clear to me that there's anything else out there.)

There are a number of folks who insist that they have to have ASM 
because
they use the PIM-RP to do source discovery.  PIM-RPs for ASM 
(specifically not
including bidir here) lead to the requirement for MSDP.  MSDP was
the right thing at the time, but it has outlived its usefulness.

We need a dynamic source discovery capability.  How it is implemented
may actually depend on the application.  It may not be necessary for
a new multicast directory protocol.  What is necessary is that the
dependency on the PIM-RP and the complicated process of switching
from the shared-tree to the source tree be deprecated so debugging
the multicast session becomes less of a problem.  ie, one should be
able to traceroute to the source (where the RPF path is followed)
and tell immediately where the break is.  With "Standard Sparse-mode 
PIM",
one can never be sure without spending an enormous amount of time
tracking down pim neighbors and other things that are just not
obvious.