Re: [Dtls-iot] Reference to mathewson-no-gmtunixtime

[Hannes Tschofenig <[e-mail]>]

Hi Carsten,

thanks for sharing your views. Good that you noticed this paragraph that
would indeed make it possible to add appropriate text to the draft.

I will work on a proposal and share it with the group.

Ciao
Hannes

On 07/15/2015 01:46 PM, Carsten Bormann wrote:
> Hannes Tschofenig wrote:
>> a) Remove the reference to mathewson-no-gmtunixtime and point out that
>> somewhere a secure time source needs to be made available.
>>
>> b) Copy the relevant text from mathewson-no-gmtunixtime  into this
>> document (with appropriate attribution). mathewson-no-gmtunixtime  cites
>> a different motivation for doing what he suggests, which I believe is
>> less applicable to our scenario.
>>
>> c) Work with Mathewson on mathewson-no-gmtunixtime  to get it finished.
>> The profile document would be blocked till that time.
>>
>>
>> If you ask me for a preference then I would probably go for (b). I am
>> not sure it is, however, in the mandate of the working group to define
>> TLS-specific functionality.
> 
> Hmm, profiling this specific field in a way that makes sense for a
> constrained implementation is well in the purview of the WG.
> 
> Also, 5246 says:
>       gmt_unix_time
>          The current time and date in standard UNIX 32-bit format
>          (seconds since the midnight starting Jan 1, 1970, UTC, ignoring
>          leap seconds) according to the sender's internal clock.  Clocks
>          are not required to be set correctly by the basic TLS protocol;
>          higher-level or application protocols may define additional
>          requirements.  Note that, for historical reasons, the data
>          element is named using GMT, the predecessor of the current
>          worldwide time base, UTC.
> 
> "Clocks are not required..."
> "higher-level or application protocols may define..."
> 
> The license to fix this in a profile is right there.
> 
> (b) is the right way to handle this. ✅
> 
> (On a technical level, it seems the intention was to have a source of
> entropy that also has a good chance to be unique over time.  No idea
> whether this sentiment should be picked up here.)
> 
> Grüße, Carsten
> 
> _______________________________________________
> dtls-iot mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/dtls-iot
>