Re: [Roll] regarding DIS Modifications draft (draft-goyal-roll-dis-modifications-00)

Allow me to reiterate the question that I asked at the mike: why is the behavior proposed by this draft (i.e., to be not considering receipt of a DIS as an indication of inconsistency and thus not resetting trickle timers) not default behavior in RPL?

That the WG during the meeting did not have a clear answer to this leads me to believe that it should be - but I wanted to ask here as there may be a subtle reason....(and, if so, the conditions in which it applies or not should be called out in any IDs modifying the behavior).

Cheers,

Thomas

-- 
Thomas Heide Clausen
http://www.thomasclausen.org/

"Payload (noun): wasted bandwidth between headers" (C.Lavenu 2011)

On 14 Nov 2011, at 19:09, Yoav Ben-Yehezkel <yoav@yitran.com> wrote:

> Thanks,
>  
> I didn’t notice the 2^ J
> I take my comment back.
>  
> Best regards,
> Yoav
>  
> From: roll-bounces@ietf.org [mailto:roll-bounces@ietf.org] On Behalf Of Emmanuel Baccelli
> Sent: Monday, November 14, 2011 4:00 AM
> To: roll@ietf.org
> Subject: Re: [Roll] regarding DIS Modifications draft (draft-goyal-roll-dis-modifications-00)
>  
> Hi Yoav,
> good call. In fact the current specification interprets the randomization range being from 0 to 2^SpreadingInterval ms. The maximum interval is thus  0 to 2^256 ms, which is quite an astronomic number already! Do you think we need some more fancy way to define the range from the number specified in the SpreadingInterval field?
> Cheers
> Emmanuel
>  
>  
> On Sun, Nov 13, 2011 at 5:32 PM, Yoav Ben-Yehezkel <yoav@yitran.com> wrote:
> Hi,
>  
> I went through the draft, it seems interesting.
>  
> I had one concern regarding the use of the Response Spreading Option specified in 4.2.
>  
> 8 bits specified in ms can be a too small randomization for some L2 technologies (such as narrowband powerline).
>  
> Would it be possible to increase the number of bits (although 16 seems like an overkill)?
> Another option could be to divide the field into two (first part units: ms, 10ms, 100ms, etc. and second part randomization range)?
>  
> I expect that in order to avoid collisions this randomization is best achieved if the units are at the magnitude of the duration of a frame or so.
>  
> Best regards,
> Yoav
>  
> 
> _______________________________________________
> Roll mailing list
> Roll@ietf.org
> https://www.ietf.org/mailman/listinfo/roll
> 
>  
> _______________________________________________
> Roll mailing list
> Roll@ietf.org
> https://www.ietf.org/mailman/listinfo/roll