All things related to multicast

[Bill Nickless <>]

At 09:59 AM 8/7/2002 -0700, Kevin C. Almeroth wrote:
> As for the value of ASM, it certainly has a niche.  The point though
> is to separate ASM from MSDP and come up with some additional options.
> 'Cause right now, ASM using MSDP stinks.  So maybe that second option
> moves the source discovery functionality out of the network.

Pick 1:

 - Equal probability of the 1st, 2nd, ... nth packet of a newly
   active source being delivered to interested ASM receivers.

 - Source-rooted distribution trees.

Lots of people can't see past the model that the Internet should be like an 
old-fashioned 10Base5 thick Ethernet segment--transmit at any time (without 
warning) and it's fine!  So any spec, implementation, or practical 
deployment of MSDP that won't achieve that (impossible) goal was thought to 
"stink".  (Yes, I think PIM Sparse Mode suffers from the same problem.)

Put another way--

ASM does not support pre-notification that a source is or will be 
active.  SSM does support pre-notification that a source is or will be 
active, but it does so at the wrong place--the interested receivers, not 
the soon-to-be-active source.

What I'm advocating is for sources to register their intent to "go active" 
so the knowledge of that intent can be spread explicitly.  Again, this is 
in contrast with the 10Base5 thick Ethernet model where you don't have to 
do anything before putting a frame on the wire.

For a more complete treatment of this issue, see

http://www.ietf.org/internet-drafts/draft-ietf-mboned-iesg-gap-analysis-00.txt



===
Bill Nickless    http://www.mcs.anl.gov/people/nickless      +1 630 252 7390
PGP:0E 0F 16 80 C5 B1 69 52 E1 44 1A A5 0E 1B 74 F7