Re: [Roll] [6tisch] Support of flow label to carry the RPL information in data packets

[Fernando Gont <[e-mail]>]

Hi, Brian,

On 04/17/2014 05:31 PM, Brian E Carpenter wrote:
>>
>> On 04/17/2014 11:27 AM, Pascal Thubert (pthubert) wrote:
>>> Considering the support we have at 6TiSCH and ROLL for the work, I
>>> published draft-thubert-6man-flow-label-for-rpl-00.txt to the 6MAN WG.
>>>
>>> The main discussion is probably to confirm whether our proposed use of
>>> the flow label inside the RPL domain is compatible with the goals that
>>> are achieved by RFC6437. Let us continue the discussion there from now on.
>>
>> I just skimmed through the I-D. Two quick questions:
>>
>> 1) Does your document propose/require that the Flow Label be rewritten
>> by some border router? -- I ask because, at the time, there was strong
>> opposition to this.
> 
> It was of course contested, but we have to remember that the flow label
> is an unauthenticated field, so our interpretation of it has to be
> robust against altered values. RFC 6437 indeed covers the case where a
> forwarding node sets the flow label.

FWIW, I was just asking to get an overall/quick understanding of where
they wanted to go, and contrast that where 6man seemed to want to go at
the time. Just that. IHMO, if the point/use case is valid, we should
"evaluate" their proposal (i.e., by no means I was meaning "6man doesn't
like that, forget it".)


>> 2) Does the algorithm/scheme with which you'll rewrite the Flow-Label
>> lead to predictable sequences? And/or, would the resulting values have a
>> uniform distribution?
> 
> The same part of RFC 6437 does RECOMMEND a uniform distribution.

Agreed. Again, I was checking the extent to which they might be
deviating from that was set here.

Thanks,
-- 
Fernando Gont
SI6 Networks
e-mail: [e-mail]
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492