On Apr 8, 2025, at 05:03, Nate Karstens wrote: > Chairs, > This draft was discussed by pim during IETF 118 & 120 and intarea during > IETF 121 & 122. There have also been some reviews on the mailing list. Would > this be an appropriate time to consider adoption by the intarea WG? > > https://datatracker.ietf.org/doc/draft-karstens-intarea-multicast-application-port/ > Regards, > Nate
Following up on this request... Can we have a few minutes of agenda time in the Internet Area Working Group meeting in Madrid to remind people about this work and initiate a call for working group adoption? In 1997 at the SIGCOMM Conference in Cannes, for the humorous “Outrageous Opinions” evening session I proposed that IP multicast should not use UDP but should instead use “MDP” (Multicast Datagram Protocol) which is like UDP but without port numbers. Of course, by then it was already too late to make such a change. Then, 27 years later, Nate Karstens proposed that IP multicast should use UDP, but with a single reserved “dummy” placeholder value where the destination port would go for a unicast packet. (If you are unclear why the UDP destination port is nonsensical for a multicast packet, please read the draft.) This is an inspired solution to what I was thinking in 1997, but done in a way that is practical and actually deployable. Nate’s proposal is: * Architecturally superior to today’s situation * Solves a real problem of port allocation * The draft is an easy read -- just two pages of actual content * The idea is simple * The idea is incrementally deployable - Applications can use it today with existing networking APIs - Future updates to APIs can add protections to prevent misuse of this reserved pseudo-port Please, let’s adopt this. Stuart Cheshire <https://datatracker.ietf.org/doc/html/draft-karstens-intarea-multicast-application-port> _______________________________________________ Int-area mailing list -- int-area@ietf.org To unsubscribe send an email to int-area-le...@ietf.org
