Re: [tram] Last Call: <draft-ietf-tram-alpn-06.txt> (Application Layer Protocol Negotiation (ALPN) labels for Session Traversal Utilities for NAT (STUN) Usages) to Proposed Standard

["Prashanth Patil (praspati)" <[e-mail]>]

Hi Spencer,
Yes, a middle box could still use ALPN identifiers to make policy decisions. But that’s true with any ALPN identifier and not specific to identifiers defined in this draft.

-Prashanth

From: Spencer Dawkins <[e-mail]<mailto:[e-mail]>>
Date: Monday, October 27, 2014 7:38 PM
To: Prashanth Patil <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, Simon Perreault <[e-mail]<mailto:[e-mail]>>
Cc: "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Subject: Re: [tram] Last Call: <draft-ietf-tram-alpn-06.txt> (Application Layer Protocol Negotiation (ALPN) labels for Session Traversal Utilities for NAT (STUN) Usages) to Proposed Standard

On 10/15/2014 02:01 AM, Prashanth Patil (praspati) wrote:

Hi Adam,
The concerning use cases described in the draft were to indicate that the new ALPN identifiers could also be considered for WebRTC firewall traversal (as pointed out in http://tools.ietf.org/html/draft-hutton-rtcweb-nat-firewall-considerations-03#section-4.2).
End to end identifier negotiation is still the goal of this draft i.e. to negotiate the use of stun and its usages.
We'll remove these use cases and stick to end to end negotiation.

Sorry for chiming in late. I saw this exchange when it happened, but I'm still thinking about it.

I understood Adam's point to be "the draft says you can use ALPN labels to do something we don't think people should do", and Prashanth's response to be "we'll make sure the draft doesn't say that".

That might be the right action.

What finally sunk in for me, was that changing the draft so that it doesn't include those use cases, doesn't prevent anyone from doing what Adam doesn't think should be done, does it?

Would it be better to point out this concern in the draft, and perhaps explain why this is discouraged?

That's not a suggestion that you should do that. It's an honest question.

Thanks,

Spencer, as your more alert AD ...


-Prashanth
> On Thu, Oct 9, 2014 at 15:41, Adam Langley <[e-mail]><mailto:[e-mail]> wrote:

>> Is your concern specific to TRAM's proposed use of ALPN, or does it apply to
>> ALPN in general?

> In general I think the IETF should be promoting the end-to-end
> principle. My concern is specifically about TRAM implicitly endorsing
> the idea that the network should be applying policy like that.

> (There's also a fair amount of irony in that fact that the second
> example suggests ALPN be used to get around the fact that networks
> often discriminate based on TCP port numbers, and the first example is
> suggesting that networks do exactly the same thing to ALPN
> identifiers.)

> Cheers

> AGL



_______________________________________________
tram mailing list
[e-mail]<mailto:[e-mail]>https://www.ietf.org/mailman/listinfo/tram