Esteban, You are right, if a mote doesn't send anything else than EBs and DIOs (which are all sent to link-layer address 0xffff), it will not increase the BE value. It is only when the starts particpating in a join procedure, or exchanges application-level data that is will realize it doesn't receive an ACK and increase BE. I do believe that the duration after being switched on that a mote only sends EBs and DIOs is short.
Thomas On Fri, Feb 10, 2017 at 12:21 PM, Esteban Municio < [email protected]> wrote: > Hi Thomas, > > No, that's exactly what I ment. > > I was just thinking about how the network would behave when for example > 20 o 30 nodes are switched on simultaneously in Minimal. If the initial > value of BE is low (1 or 2), collisions are expected to be very frequent > in the shared cell. And since the first messages for joining the network > (EBs and DIOs) are broadcasted without expecting an ACK, the BE could > remain unmodified and the joining process could last forever. > > Could this scenario happen? Or am I missing something important? > > But indeed maybe this is out of the scope of Minimal since is just > matter of sizing the number of shared cells in the schedule and the > topic is rather related with the 802.15.4 standard than with 6tisch. > > Kind regards, > Esteban > > > On Thu, 2017-02-09 at 23:46 +0000, Thomas Watteyne wrote: > > Esteban, > > > > > > I believe this is fully specified in the 802.15.4 standard. In a > > nutshell, the back-off mechanism is triggered only when the cell is > > marked as "shared". When a mote sends a packet and doesn't get an ACK, > > the BE is changed. If you send a frame to link-layer destination > > 0xffff, you're not expecting an ACK, so BE will not change. > > > > > > I may be missing your point? > > > > > > Thomas > > > > On Thu, Feb 9, 2017 at 10:59 AM, Esteban Municio > > <[email protected]> wrote: > > Hi all, > > > > I was wondering if there is any recommendation about how set > > up the > > Backoff mechanism in shared cells when no ACKs are expected. > > > > According to the standard, macMaxBE and macMinBE are defined > > and the BE > > is increased at every attempt. However, if the node is not > > expecting any > > ACK (i.e sending DIOs or EBs), I guess the BE will be always > > the same. > > Is there any recommendation in Minimal about how to choose > > this value? > > > > Maybe a value related with the number of nodes or DIO/EB > > period? > > > > Kind regards, > > Esteban > > > > -- > > Esteban Municio > > IDLab - Dept. Mathematics and Computer Science > > University of Antwerp - imec > > Middelheimlaan 1 , 2020 Antwerp, Belgium > > Office M.G.325 > > > > _______________________________________________ > > 6tisch mailing list > > [email protected] > > https://www.ietf.org/mailman/listinfo/6tisch > > > > > > > > > > > > -- > Esteban Municio > IDLab - Dept. Mathematics and Computer Science > University of Antwerp - imec > Middelheimlaan 1 , 2020 Antwerp, Belgium > Office M.G.325 > -- _______________________________________ Thomas Watteyne, PhD Research Scientist & Innovator, Inria Sr Networking Design Eng, Linear Tech Founder & co-lead, UC Berkeley OpenWSN Co-chair, IETF 6TiSCH www.thomaswatteyne.com _______________________________________
_______________________________________________ 6tisch mailing list [email protected] https://www.ietf.org/mailman/listinfo/6tisch
