headers
Pasi Sarolahti | 17 May 2011 14:20
Picon
Picon
Favicon
Gravatar

Fwd: I-D Action: draft-ietf-dccp-udpencap-08.txt

Hi,

The udpencap draft is updated based on the input from the MMUSIC group. The changes concern the SDP part in
the draft, and the other parts seem unchanged. Thanks to Colin and Gorry for the update!

Based on the earlier concluded WGLC, I assume the other parts are still ok with the WG, unless somebody says
otherwise. I hope people can check the new SDP text, and send any comments soon.

PS. I see that the draft is also waiting for an ICE example. What are your plans with that?

- Pasi

Begin forwarded message:

> From: <internet-drafts <at> ietf.org>
> Date: May 17, 2011 10:34:54 AM GMT+03:00
> To: <i-d-announce <at> ietf.org>
> Cc: <dccp <at> ietf.org>
> Subject: [dccp] I-D Action: draft-ietf-dccp-udpencap-08.txt
> 
> A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work
item of the Datagram Congestion Control Protocol Working Group of the IETF.
> 
> 	Title           : Datagram Congestion Control Protocol (DCCP) Encapsulation for NAT Traversal (DCCP-UDP)
> 	Author(s)       : Tom Phelan
>                          Godred Fairhurst
>                          Colin Perkins
> 	Filename        : draft-ietf-dccp-udpencap-08.txt
> 	Pages           : 19
> 	Date            : 2011-05-17
(Continue reading)

Permalink | Reply |
headers
Rémi Denis-Courmont | 17 May 2011 14:29
Favicon

Re: Fwd: I-D Action: draft-ietf-dccp-udpencap-08.txt

   Hello,

On Tue, 17 May 2011 15:20:30 +0300, Pasi Sarolahti <pasi.sarolahti <at> iki.fi>
wrote:
> Based on the earlier concluded WGLC, I assume the other parts are still
ok
> with the WG, unless somebody says otherwise. I hope people can check the
> new SDP text, and send any comments soon.

There are still absolute zero considerations for programming interfaces.
And with it, my concern that the duplicated pairs of port numbers will make
it hard if at all possible to integrate in existing applications. VLC won't
be able to implement this extension to DCCP and I really wonder if anyone
actually uses it.

--

-- 
Rémi Denis-Courmont
http://www.remlab.net/
Permalink | Reply |
headers
Colin Perkins | 17 May 2011 22:56

Re: Fwd: I-D Action: draft-ietf-dccp-udpencap-08.txt

On 17 May 2011, at 13:29, Rémi Denis-Courmont wrote:
> On Tue, 17 May 2011 15:20:30 +0300, Pasi Sarolahti <pasi.sarolahti <at> iki.fi> wrote:
>> Based on the earlier concluded WGLC, I assume the other parts are still ok with the WG, unless somebody
says otherwise. I hope people can check the new SDP text, and send any comments soon.
> 
> There are still absolute zero considerations for programming interfaces. And with it, my concern that
the duplicated pairs of port numbers will make it hard if at all possible to integrate in existing
applications. VLC won't be able to implement this extension to DCCP and I really wonder if anyone actually
uses it.

What's the concern here? Use the IANA registered port, unless specified otherwise by the application. Any
UDP tunnelling solutions must specify a UDP port.

--

-- 
Colin Perkins
http://csperkins.org/
Permalink | Reply |
headers
Rémi Denis-Courmont | 18 May 2011 07:22
Favicon

Re: Fwd: I-D Action: draft-ietf-dccp-udpencap-08.txt

On Tue, 17 May 2011 21:56:52 +0100, Colin Perkins <csp <at> csperkins.org>
wrote:
> What's the concern here? Use the IANA registered port, unless specified
> otherwise by the application. Any UDP tunnelling solutions must specify
a
> UDP port.

The concern is that we have two (pairs of) ports. This does not only not
fit in the standard SDP m=line, but it does not fit in the traditional
sockaddr_in/sockaddr_in6 abstraction (and its equivalent in many
programming languages/frameworks).

It might be that the API makes the problem go away, but that seems like a
risky bet in the absence of any sketch of an API (it would not need to be
normative).

--

-- 
Rémi Denis-Courmont
http://www.remlab.net/
Permalink | Reply |
headers
Colin Perkins | 18 May 2011 08:24

Re: Fwd: I-D Action: draft-ietf-dccp-udpencap-08.txt

Rémi,

On 18 May 2011, at 06:22, Rémi Denis-Courmont wrote:
> On Tue, 17 May 2011 21:56:52 +0100, Colin Perkins <csp <at> csperkins.org>
> wrote:
>> What's the concern here? Use the IANA registered port, unless specified otherwise by the application.
Any UDP tunnelling solutions must specify a UDP port.
> 
> The concern is that we have two (pairs of) ports. This does not only not fit in the standard SDP m=line, but it
does not fit in the traditional sockaddr_in/sockaddr_in6 abstraction (and its equivalent in many
programming languages/frameworks).

How do you suggest we build a UDP encapsulation solution without using a second pair of ports? It would seem
fundamental to the solution space.

> It might be that the API makes the problem go away, but that seems like a risky bet in the absence of any sketch
of an API (it would not need to be normative).

Do you want to propose an API? 

--

-- 
Colin Perkins
http://csperkins.org/
Permalink | Reply |
headers
Rémi Denis-Courmont | 19 May 2011 13:54
Favicon

Re: Fwd: I-D Action: draft-ietf-dccp-udpencap-08.txt

On Wed, 18 May 2011 07:24:19 +0100, Colin Perkins <csp <at> csperkins.org>
wrote:
>> It might be that the API makes the problem go away, but that seems like
>> a risky bet in the absence of any sketch of an API (it would not need
to
>> be normative).

> Do you want to propose an API?

I have no _sane_ ideas how to build an API that works with the current
proposal, that's why I'm concerned.

--

-- 
Rémi Denis-Courmont
http://www.remlab.net/
Permalink | Reply |
headers
Colin Perkins | 21 May 2011 14:00

Re: Fwd: I-D Action: draft-ietf-dccp-udpencap-08.txt

On 19 May 2011, at 12:54, Rémi Denis-Courmont wrote:
> On Wed, 18 May 2011 07:24:19 +0100, Colin Perkins <csp <at> csperkins.org> wrote:
>>> It might be that the API makes the problem go away, but that seems like a risky bet in the absence of any
sketch of an API (it would not need to be normative).
> 
>> Do you want to propose an API?
> 
> I have no _sane_ ideas how to build an API that works with the current proposal, that's why I'm concerned.

What do you need other than one new socket option to indicate that UDP encapsulation is to be used and to
provide the port? I must be missing something, since the API doesn't seem difficult. 

--

-- 
Colin Perkins
http://csperkins.org/
Permalink | Reply |
headers
Colin Perkins | 17 May 2011 22:54

Re: Fwd: I-D Action: draft-ietf-dccp-udpencap-08.txt

On 17 May 2011, at 13:20, Pasi Sarolahti wrote:
> Hi,
> 
> The udpencap draft is updated based on the input from the MMUSIC group. The changes concern the SDP part in
the draft, and the other parts seem unchanged. Thanks to Colin and Gorry for the update!
> 
> Based on the earlier concluded WGLC, I assume the other parts are still ok with the WG, unless somebody says
otherwise. I hope people can check the new SDP text, and send any comments soon.
> 
> PS. I see that the draft is also waiting for an ICE example. What are your plans with that?

If there's agreement that the approach makes sense, them I'll write an example. 

--

-- 
Colin Perkins
http://csperkins.org/
Permalink | Reply |

Gmane