Re: proposed charter

In message <[email protected]>
Fred Baker writes:
>  
>  
> On Sep 27, 2006, at 11:44 AM, Curtis Villamizar wrote:
>  
> > Just confirming that Fred considers ETS to be voice only - at least
> > for now.  btw - I think that is a valid assumption - at least for now.
>  
> depends on which "ETS" you're talking about. In the military space,  
> current proposals suggest at least one "preferred elastic" class  
> similar to AF for elastic applications as well as preferential real  
> time services for voice and video. In the civilian space, unless  
> Janet-et-al tell me otherwise, the current discussion is of an  
> Internet counterpart for GETS, which is to say that it is about voice.


In the military space that I've heard about there is no non-military
traffic so plain Ol' EF and AF should do it.  If the military is using
or begins using the commercial Internet as part of their ETS, then
they can request DSCP codepoints for their EF (more than one is never
needed if traffic is policed) and as many AF classes as they need
(within reason).  This would give military and non-military traffic
separate DSCP codepoints to work with.  Again, this might be a "later"
when the military has deployed EF and AF and needs to extend this over
the commercial Internet.  If it is commercial Internet from day one,
they can still use an existing AF class as long as the provider(s)
have one free.  It is very easy to do a DSCP mapping at an AS boundary
and DSCP marked traffic is something exchanged only "among consenting
adults".  Such a mapping would make a smooth transition possible.

Some of the things the military has pushed for either didn't happen or
didn't work.  For example, MARS and NHRP.  It would be premature to
assign codepoints for hypotheticals.

So I think the "just adding one EF codepoint for ETS is OK for now"
assumption still holds with the emphasis on "at least for now".

Curtis