Minutes for HOMENET at IETF-96
minutes-96-homenet-2
Meeting Minutes | Home Networking (homenet) WG | |
---|---|---|
Date and time | 2016-07-18 12:00 | |
Title | Minutes for HOMENET at IETF-96 | |
State | Active | |
Other versions | plain text | |
Last updated | 2016-08-01 |
minutes-96-homenet-2
IETF 96 - Homenet Agenda Monday, July 18th, 2016 14:00-15:30 Afternoon Session I Potsdam I 0. Administrivia (5m) Blue Sheets Note taker - Tim Wicinski Jabber relay - Barbara Stark 1. WG Status Update - Chairs (5m) Documents Published since last meeting - RFC 7787 - Distributed Node Consensus Protocol - RFC 7788 - Home Networking Control Protocol Updated Drafts: draft-lemon-homenet-naming-architecture-01 draft-cheshire-homenet-dot-home-03 New Drafts: - draft-ietf-homenet-babel-profile-00 (Adopted by WG) Drafts on hold pending other docs: - draft-ietf-homenet-hybrid-proxy-zeroconf-02 - draft-ietf-homenet-front-end-naming-delegation-04 - draft-ietf-homenet-naming-architecture-dhc-options-03 2. Routing HNCP Deployment Experiences - Juliusz Chroboczek (15m) JC: HNCP is designed to configure unmanaged, small, stable, prefix-based networks Feel that can be pushed beyond this. Built a testbed for Babel. hnetd works with small sample size hnetd has limitations with bugs, undocumented, and missing features hnetd is mising some features prefix announcement static prefix (unknown, implementor): Did not talk about Babel, works well? JC: If Babel break in this network, we fix it first. Pierre Pfister(PP): Did you use the ad-hoc mode JC: Used ad-hoc, no need to renumber #2: /80 thing, use SLAAC, will fight for soft subnet boundary. can not do NAT with a 64 bit size. soft subnet sizes are dangerous JC: Agree with /64, but at same time wish to subnet /64 Mikael Abrahamsson: why the /62 space JC: Pure accident of history Dave That(DT): Work on Android at this point? JC: Lorenzo may answer that better Lorenzo: Power important problem to solve, if sending/rcv packets every 3 seconds DT: Tethering scenario (unknown): Interested in finding code JC: All code is part of OpenWRT, except for tcpdump modifications, which are being worked on. 3. Naming Architecture and Service Discovery Architecture Draft - Ted Lemon (15m) - draft-lemon-homenet-naming-architecture-01 Ted Lemon(TL): New version out, removed a lot of text. Andrew Sullivan(AS): This is good enough to adopt and work on as a working group. Let's get a draft and start working on it. TL: Hybrid mDNS/DNSSD: problems name conflict, no security, name collisions Proposed Solution: secure way forward, use mDNS to defend names, homenet devices do not advertise using mDNS non-homenet aware DNSSD devices still use mDNS hostname conflicts have link name attached need a document to explain the zone signing should develop a standard API for managing devices Documents: mds and dnssd work how dns works on homenet dns on interacts with outside world global names are acquired JC: Read document as strawman, but much better written. use local resolver, which shncpd can not do. ula issue TL: Something on local link as resolver JC: understanding is single resolver in homenet TL: reason want on every link, if request is to link-local, then link-local query PP: mDNS/DNS Proxy, mDNS was there. can we modify the host? TL: Describe what the host could implement PP: If the host does not implement? TL: Still works using mDNS hybrid AS: Step up a level, if you don't have that ambition, you will not get the things you want with homenet DT: Possible to stage in multiple pieces? Chairs: take that to the list Chairs: Hum to adopt to as WG item *ITEM*: Hum to adopt was strong, no hum against RFC7788-bis and ".home" - Chairs / AD / Jari Arkko (50m) - draft-ietf-homenet-hncp-bis-00 Three options: 1 - new RFC removed all mention of .home 2 - new RFC froming all DNS or naming to .home 3 - - new RFC that explains procedural failure the occurred and no long identifies .home Terry Manderson(TM): made it all the way through the process and failed Jari Akro(JA): Ways forward. Formal RFC issued with errata filed rather quickly. put the name in a separate step. names like ".home" are special use names, technical problems around this. TL: if we remove .home, something should be put in there. TM: We need to address this problem right away. Tim Wicinski(TW): remove the .home, but also the bis addresses the mDNS issues. Mark Townsley(MT): Chairs prefer #3 Stuart Cheshire(SC): the 'fix' is like some legal technicality. .home draft got little discussion Steven Barth(SB): errata means name depends on whatever vendor uses. TM: Parallel tracks are OK PP: Asked to integrate errata, did not want to. prefer the third solution Ralph Droms(RD): Needs to think what a .home resolution needs to do. Suzanne Woolf (SW): Understand frustration with process. Actively working on it. Thanks to Homenet for providing the forcing function. Lee Howard(LH): 7788 not deployable with .home as ultimate name is used. Dan York(DY): do not like any of the options. Implementer implementing homenet don't care of our process failure. James (J): Implementer, would like to implement what is out there Option 5: Don't do anything. Option 1: no hum Option 2: light hum option 3: hum option 4: hum option 5: light hum option 6: (something else we have not thought of) - strong hum