ietf
[Top] [All Lists]

Re: Multicast

2001-03-07 04:20:04

In message <3AA60F02(_dot_)2FE41239(_at_)irisa(_dot_)fr>, Ali Boudani typed:

Isnt SSM just a particular case of PIM??

the right place for this discussion is
  SSM <ssm-interest(_at_)external(_dot_)cisco(_dot_)com>, 

SSM is a subset of PIM SM, roughly, and relies (sort of) on IGMPv3 (at
least on a subset of equiv. functionality).

It is the specifications for just specific sources but they arent adressing
the multicast in general.
am I right ?

not quite - i dont think the whole IETF list is the right place for
this one - see
http://www.ietf.org/html.charters/ssm-charter.html

i think since the ssm work is close to done, we'll see work resume on
bidir (s/cbt/pim-bidir:-) soon , and similalrly in the RMT work
see
http://www.ietf.org/html.charters/rmt-charter.html
a lot of building blocks are close to done (prob. about a year) then
we'll see some work on multiple source (the latent demand for multiple
source applications is imho underestimated, but until we can fix the
more immediate problems with supporting 1-many, we can't really expect
people to deploy many-to-many extensively- other problems being
addressed are concerned with having good solutions for multicast security (see 
http://www.ietf.org/html.charters/msec-charter.html
and for many-to-many, for congestion control (to meet transport area
requirements)....

i think (but of course i am usually wrong) that we may see progress on
this in 2002...

Jon Crowcroft wrote:

In message <3AA60C4B(_dot_)58AA4CFC(_at_)irisa(_dot_)fr>, Ali Boudani 
typed:

 >>First the CBT protocol was created to use shared tree solutions because
 >>DVMRP and the other dense mode protocols werent scalable. there were
 >>many problems with CBT (which is bidirectional) so PIM-SM was cretaed
 >>which provide some switching (between shared tree and source tree). and
 >>after that there is some discussions about the bidirectional PIM, which
 >>is like CBT.
 >>Are we in circle here or what ??

not really. the mainstream current multicast action is concentrating on
 single source (and on single source reliable multicast transport)
since we didn't feel we understood all the complications of ANY of the
multiple source schemes for IP or reliable....(e.g. interdomain
routing, and multiple source semantics for reliable) -

there were'nt really "problems" with CBT apart from we never managed
to get a router vendor to committ to an implmenetation which we could
deploy and learn from - tony ballardie got a lot of the details out,
but the two implementaions i know of never saw light of day.....bidir
pim is cool, bgmp is cool, but action in implementation/details/spec
is waiting on getting the PIM SSM stuff completely shaken down....

its all part of a good learning experience and (as any good s/w
engineer might say) its the norm:-)
 >>

 cheers

   jon


 cheers

   jon



<Prev in Thread] Current Thread [Next in Thread>