Re: [dnssd] DNS Name Autoconfiguration for Home Network Devices

The term of "IoT device" is mentioned many times here and in the draft, and it seems necessary to give a relatively formal defination to it instead of only presenting some instances of so called "IoT device". The domain name is usually used to uniquely and consistently identify objects and thus it is better to keep the domain name unchangeable. If many kinds of information of "IoT devices" (such as their locations and categories) are inserted into the domain name in some way, whether we should change the domain name or not if some kinds of information (like the location) changes? 

Guangqing Deng
cnnic 

From: Mr. Jaehoon Paul Jeong
Date: 2014-11-19 12:00
To: Stuart Cheshire
CC: dnssd@ietf.org; Myung-Ki Shin; Brian Haberman; Jung-Soo Park
Subject: Re: [dnssd] DNS Name Autoconfiguration for Home Network Devices
Stuart,
Thanks for your constructive comments.

As I answered Alf's email message just before,
I would like to allow IoT vendors to have an alternative way for DNS naming
with more compact implementation based on IPv6 stateless autoconfiguration.

For the question about how users find out the host names and which protocols are used
were presented with the following slides 4 and 5 during the last dnssd WG session in Hawaii:
http://www.ietf.org/proceedings/91/slides/slides-91-dnssd-6.pdf

To figure out what those devices do, we can use device category and device model 
in the DNS name format.

Paul

===========================
Mr. Jaehoon (Paul) Jeong, Ph.D.
Assistant Professor
Department of Software /
Department of Computer Engineering
Sungkyunkwan University
Office: +82-31-299-4957
Mobile: +82-10-4758-1765
Fax: +82-31-290-5119
Email: pauljeong@skku.edu, jaehoon.paul@gmail.com
CPS Lab Website: http://cpslab.skku.edu
Personal Homepage: http://cpslab.skku.edu/people-jaehoon-jeong.php

On Wed, Nov 19, 2014 at 7:46 AM, Stuart Cheshire <cheshire@apple.com> wrote:

On 17 Nov, 2014, at 17:33, Mr. Jaehoon Paul Jeong <jaehoon.paul@gmail.com> wrote:

> Hi Tom,
>
> You are right for appliances with enough capacity to run mDNS.
>
> However, for low capacity IoT devices without mDNS, my proposal will be able to support an alternative way for DNS naming.

You assert that there are devices with insufficient resources to run DNS-SD/mDNS.

You assert that these devices would have sufficient resources to run your alternative.

I think you need data to substantiate these claims. Note that devices like the SitePlayer Telnet implement DNS-SD/mDNS in under one kilobyte (about 850 bytes of code, if I remember correctly): <http://netmedia.com/siteplayer/telnet/>

So you need to demonstrate that:

1. That there are devices where 850 bytes of code is too much.

2. That these devices can implement your thing in under 850 bytes.

3. That your thing is useful. I’m still unclear on what your thing is useful for. Your proposal says how devices pick host names, but not how users find out what those host names are, or find out what those devices do, or what protocol to use to communicate with them, and how clients resolve those host names to addresses.

Stuart Cheshire

_______________________________________________
dnssd mailing list
dnssd@ietf.org
https://www.ietf.org/mailman/listinfo/dnssd