Re: Use of SAP in a PIM SSM Network

[Hugh LaMaster <lamaster@nren.nasa.gov>]

On Wed, 15 Nov 2000, Jon Crowcroft wrote:

> ha ha ha
> 
> finally people realize
> 
> ssm entails _exactly_ as many problems as simple multicast in terms of
> 1/ deploying it in routers
> 2/ changing all the applications to use igmpv3
> 3/ chaing host OS to use igmpv3
> 4/ changing session management/annoucement mechanisms
> 
> :-)

I don't know how many people agree with me, but, although
SSM obviously has huge benefits for broadcast-type applications,
the benefits are not so obvious to me for all multicast applications.
So, personally, although I fully support the SSM deployment,
I don't want to abandon completely the capability for global 
shared trees, which I think will continue to be most appropriate
for large numbers of low datarate sources, such as distributed
simulations, as well as multicast-enabled games that were 
"threatened" a few years ago.

At the moment, the assignment of the SSM address space and
a wave of deployments over the next year will give us a
usable SSM infrastructure - that is, addressing item 1/
above - although, of course, /2, 3/, and 4/ will still be
there.  /4 is solvable, though not solved yet.  2/ and 3/ 
are what I see as the main problem, and even after solutions
are available, still won't be *automatic* in the case where 
that matters most -- "personal systems" (e.g. the gaggle of
PC-based systems running legacy Windows 95/98/me).

SSM does seem to have finally given multicast the appeal to 
commercial content providers that was not there before, though, 
so that demand-pull from users may finally start to have an
effect.

> so no, we dont have resources to do this, but we are happy to roll in
> SSM/igmpv3 changes to the mbone apps if people contrib. them (spritn
> have kindly offered to do some apps) - sdr is a non trivial one in
> that it needs completely re-writing which is outside the scope of a
> university research department's normal activities....

If somebody is going to rewrite sdr to integrate SSM, I hope 
it is done so that it still handles traditional SAP announcements 
transparently in parallel, both on the receiving and announcement
side.  SAP will continue to be used for some time at least.


Questions/comments/refutations welcome.


--
 Hugh LaMaster, M/S 233-21,    Email: lamaster@nas.nasa.gov
 NASA Ames Research Center     Or:    lamaster@nren.nasa.gov
 Moffett Field, CA 94035-1000  Or:    lamaster@kinkajou.arc.nasa.gov
 Phone: 650/604-1056           Disc:  Unofficial, personal *opinion*.