Re: [Dtls-iot] TLS False Start ... Re: WGLC for DTLS Profile draft: 11/11/2014 - 11/25/2014

[Hannes Tschofenig <[e-mail]>]

Hi Nikos,

when I attend the TLS interim meeting in Paris I got a different
perception about the status of the TLS False Start work.

First, I was told that TLS False Start is widely deployed.

Second, the TLS chairs asked Bodo to resubmit his document and to push
it through the standardization process.

Here is a recent mail to the TLS mailing list on that topic:
http://www.ietf.org/mail-archive/web/tls/current/msg14530.html

Here is the recently submitted draft:
http://tools.ietf.org/html/draft-bmoeller-tls-falsestart-01

Finally, the text I propose for the DTLS profile draft recommends its
usage but does not mandate the implementation nor the deployment.

I did, however, had questions when reading through the draft in terms of
applicability for the IoT context, which I raised on the TLS mailing
list yesterday:
http://www.ietf.org/mail-archive/web/tls/current/msg14787.html

Ciao
Hannes

On 11/28/2014 09:38 PM, Nikos Mavrogiannopoulos wrote:
> On Fri, 2014-11-28 at 20:18 +0100, Hannes Tschofenig wrote:
>> Hi Alexey,
>>
>> On 11/28/2014 01:07 PM, Hannes Tschofenig wrote:
>>>> In 5.1: is use of something like False Start common in IoT environment?
>>>>> As it reduces number of round trips, it might be advantageous and you
>>>>> should consider mentioning it.
>>> Good question. I added it to the issue tracker:
>>> https://tools.ietf.org/wg/dice/trac/ticket/20
>> I believe TLS False Start would be a good fit. I added a section about
>> it into the draft. Here is the text:
>>    Based on the improvement over a full roundtrip for the full TLS/DTLS
>>    exchange this specification RECOMMENDS the use of the TLS False Start
>>    mechanism when clients send application data first.
> 
> TLS false start is only described in an expired draft since 2010. Apart
> from taking the TLS security guarantees to its limits (attacks that may
> require significant online effort on TLS can be converted to off-line
> attacks with false start), and having very little review from crypto
> community, there is no guarantee that an IETF false start (or even an
> informational RFC of the same document) would be compatible with it. So
> at least a more stable reference should available before recommending it
> in a standard's track document.
> 
> regards,
> Nikos
> 
> 
> _______________________________________________
> dtls-iot mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/dtls-iot
>