IETF 110 - DNSSD + Homenet

12:00-14:00 UTC Friday session I (March 12, 2021)

12:00-12:45 DNSSD

Ted Lemon presented summary of current work in DNS-SD

Slides: Combined Slides

Used slides 8-12 for discussion of DNSSD WG items.

Éric Vyncke asked about how SRP Replication will work. Ted explained why he decided not to use the normal DNS zone transfer mechanism The normal DNS zone transfer mechanism assumes a distinction between primary and secondary. SRP Replication is more peer-to-peer. The focus will not be a complete 100% robust replication protocol with transactions, rather something that is easy/compact to implement in home devices.

Ted will update draft-ietf-dnssd-srp one more time and then we can start Working Group Last Call.

We need more discussion of draft-sctl-advertising-proxy on the list, and then we can ask about Working Group adoption.

The SRP Replication draft awaits submission.

12:45-14:00 Homenet

Administrivia and Status (Chairs)

Agenda bashing took place, and it was agreed to start with the naming architecture and delegation drafts.

The Homenet Babel profile is still not published because it is waiting on the BAbel Source Specific draft. That is likely to be rewritten and go back through WGLC and IETF LC. This will take some time.

Daniel Migault / Michael Richardson (30 minutes)

Slides: doi and associated dhcp options

There was no technical discussion of these drafts and administrative discussion (how to proceed with them) was deferred to end of session.

Ted Lemon (15 minutes)

Stuart Cheshire thanked Ted Lemon for his work over the last two years, not only designing solutions (SRP and Self-configuring Stub Networks) and authoring drafts, but also writing the code to implement those solutions. This work lead to Apple shipping the HomePod mini with Thread Border Router functionality in November 2020, as well as a variety of third-party vendors shipping compatible HomeKit accessories that run over Thread using Ted’s SRP and Stub Network solutions.

Ole Trøan asked: Is “full mesh routing” just “routing”? Ted uses the term “full mesh routing” to mean multiple hops in the home across an arbitrary topology. The “stub network routing” proposal is an intentional simplification, where there is a single “backbone” home network with one or more self-configuring (permissionless) stub networks hanging off that backbone. That simplification is what makes it possible to add a new self-configuring stub network to an existing home network without requiring that other devices on the existing home network (including the home gateway and Wi-Fi access point) already know about stub networks.

Administrative Discussion -- How and Where to Proceed

There was discussion of finding the right home in the IETF to progress the stub network work.

Barbara Stark did a poll asking “Are you interested in working on stub networks?” Out of 35 people in the meeting, 12 responded “yes” and 1 responded “no”.

Barbara Stark did a poll asking “Are you willing to review draft-ietf-homenet-front-end-naming-delegation for Working Group Last Call?” Out of 35 people in the meeting, 9 responded “yes” and 1 responded “no”.

Daniel Migault will update the draft one more time and then we will begin Working Group Last Call.

Plan is to schedule a HOMENET interim meeting in about six weeks for discussion of stub networks and naming delegation. There will be a Doodle poll to select a time.