Minutes IETF99: dnssd
minutes-99-dnssd-00

DNSSD WG

IETF99, Prague
Wednesday 19th July 2017
Athens/Barcelina room 15:20 - 16:50 local time

Chairs' Introduction                                                       
Chairs, 5 mins

Status:

The dnssd-hybrid draft (now "Discovery Proxy") will be submitted to the IESG
soon. The withdrawn IPR statement on the draft needs to be clarified. The .home
reference needs to be updated to .home.arpa.

Ralph Droms is stepping down as co-chair. Looking for a co-chair.
The WG thanked Ralph for his excellent contributions to the WG as co-chair.

DNS-SD Privacy drafts                                                Christian
Huitema, 20 mins https://tools.ietf.org/html/draft-ietf-dnssd-privacy-02

We had detailed WGLC reviews by
- Stephane
- Ted

Issue; Use of PSK

Carsten Boreman (CB): Need secret handshake
Part of same category

Ted Lemon (TL):  Concerned about not load on server but caching behavior of
dnssd client.
    move from 5 minutes to 30 minutes
Unicast - 5 minutes is fine
multicast - want longer

Aanchal Malhotra (AM):  counter or abs time, counter would be better
CH: can not implement
Threat Model?
AM:  5 minutes
AM: Clock drift? accepts both in a small interval

CH: Counter

DT: pulling out into privacy or implementation considerations

SC: Did not understand the comments from AM - did a lot of work in NTP

https://tools.ietf.org/html/draft-ietf-dnssd-pairing-02

split into two drafts
- information analysis
- std track protocol

Privacy Draft passed WGLC, rady
Pairing Draft passed WGLC but
 - split analysis spec and QR code
 - need second WGLC for pairing

TL: did not read sections 2/3, easy to i

Kerry Lynn (KL): dynamic QR codes for network code.

Daniel Kaiser: Likes to split drafts

Split into Two documents:
    10 for - 1 against.  Will check on list.

QR in separate draft?

DT: three ways
- leave as is
- split into separate documents
- separate chapter

7-1-1 for a separate chapter

DNS-SD Update and New Work Items                                Stuart
Cheshire, 30 mins https://tools.ietf.org/html/draft-cheshire-dnssd-roadmap-00

why non-multicast?

all waiting on session-signaling

roadmap

client-config

https://tools.ietf.org/html/draft-sctl-service-registration-00

basically dns update

David S (DS): fcfs security and naming
SC: Some manual removal of d

MA: make simple to user
SC: key get erased. concern noted

DT: how to scale to enterprise/campus. If you haven't heard in some time,
unicast to orignal address asking. otherwise garbage collection

Juliusz Chroboczek (JC):  Can not be implemented if i do not have write able
persistent storage SC: Reg Srvr remember key until reboot. then FCFS

RD: should be documented in draft

David Schinazi(DS): what do you do with the "David's Macbook1" ?

KL: Persistent identifier that might expose info.

https://tools.ietf.org/html/draft-sctl-discovery-broker-00

TP: diff between discovery proxy vs discovery broker.
proxy is 1:1 unicast to multicast
broker is !:Many unicast to unicast

https://tools.ietf.org/html/draft-cheshire-edns0-owner-option-01

DNS-SD Advertising Proxy and Zone Stitching drafts become unnecessary as we
move away from Multicast DNS

Toke described his ddns registration implementation

Multicast DNS Discovery Proxy                                        Ted Lemon,
15 mins https://tools.ietf.org/html/draft-sctl-dnssd-mdns-relay-00

TL:

DNS-SD / CoRE RD interoperability discussion                        Kerry Lynn,
15 mins https://tools.ietf.org/html/draft-ietf-core-rd-dns-sd-00

Resource Discovery and Service Discovery interop

RD and SD are complementary; RD typically finer-grained
RD work goes back about 6 years.
CoRE WG originally intended to work with dnssd, but decided to go in its own
direction. Main output of CoRE is the CoAP protocol Multicast RESTful
applications are possible. Send query, get back set of links that match the
query. if = interface specification rt = resource type This is what we'd
probably want to export to the SD world, from the RD; so that's what the draft
discusses

SC: in that example, the rt, is that registered in iana service type registry?

CB: two registries, entries under if= and rt=

DT: potential problem is rt values could have dots in them.
SC: could in principle escape them; iana service type registry says letters,
digits, hyphens.

DT: procedural issue with kerry's example - how safe is it to assume that left
most label is special K: mapping agent could handle, or have a convention.

SC: Service is entity, has a protocol. Carsten talks about a Resource, it's
entity you talk to, it has a protocol. Sounds same! So would help to figure out
why resource is not a service.

CB: Answer is reason we have the means to identify Resources that are worth it

Summary of actions and close                                        Chairs, 5
mins

The chairs stated that with only a few people having read the new drafts by
Stuart and Ted it was too early to consider WG adoption.

1) draft-ietf-dnssd-privacy-02 will be revved from WG feedback to -03 and sent
to IESG; a document shepherd is needed

2) Confirm the actions on splitting draft-ietf-dnssd-pairing-02 on the list,
and submit updated draft(s) to include WG feedback

3) Progress discussion of Stuart and Ted's new drafts on the mail list.

4) The dnssd WG chairs will discuss a way forward for the DNS Session
Signalling draft with the dnsop WG chairs

5) Our AD will solicit expressions of interest for a new co-chair.

Ralph thanked everyone.

Meeting closed at 16:57.