Re: Work Proposal

[Jon Crowcroft <J.Crowcroft@cs.ucl.ac.uk>]

In message <3742E313.4760@cisco.com>, John Meylor typed:

 >>Jon, please.  I said we should consider ISP support on this issue
 >>carefully.  

yes, the list was announced as this type of input being task 0

 >>forward from concept to deployment is to make sure ISPs are involved.
 >>This was the key to moving forward last year, it will continue to
 >>be the key.  Randy replied, I appreciate his input.

ok....sure...and i have said time and again that this initiative is to
enhance services not detract from current deployment speed. but the
work specifically addresses creating better incentives to deployment
,and does not involve a flag day for replacing PIM for example = nor
does it in any way i see act as a disincentive for deploying hat
exists now (the argument that  an ISP would defer deploying multicast
because a better multicast may come along doesn't hold much water

 >>  My point was that if we are at a point where we are proposing
 >>new multicast service models which effect all providers in general, 
 >>we should get broad ISP input.  I hope this is not a tough or offensive
 >>concept for anyone.

yes, we are doing that. thats wghy the list exists. we already said it
several times as a pre-requisite. the routing area directors suggested
it a pre-requisite of even considering a WG proposal. I guess i should
recall that not everyone was on the list for all the discussion so
far....

 >>  Please understand, I am very supportive of this effort and new
 >>ideas in general.
 
thanks what we are trying to do now is position the work - basically
it seems silly to add items to the PIM WG - they have a very clear and
well defined, scoped and feasible set of tasks. IDMR is winding down,
and has a history of being somewhat unfocussed (its winddown is being
aschieved in a focussed way though!); so i see that if we get all the
other pieces in place, a new WG is probably appropriate. 

but we are looking at timescales of delivering a spec near end of 2000 -
most ISPs would have current style multiast out there if they were
ever going to do it by then, so this would then be a delta enabling
some things to operate better - given the end system requirement for
SM and Express, we are not going to make a windows 2000 release or
even a 2001, so my guess is that we are looking at a very much
tertiary stage of deployment of multicast....(1st was dvmrp+tnnels,
second is pim/bgmp etc)....so i think its sustainable

Dear all - ISP input PLEASE!



 cheers

   jon