Re: [6tisch] The channel hopping scheme seems to be suboptimal

[Tero Kivinen <[e-mail]>]

Dale R. Worley writes:
> Thomas Watteyne <[e-mail]> writes:
> > You raise a valid point. There are a couple of points, though, that make an
> > "Absolute Slot Number" more favorable compared to an "Absolute Slotframe
> > Number":
> > - if you schedule multiple cells between two nodes in a single slotframe,
> > you want those different transmissions to happen at a different frequency
> 
> Although you could construct the schedule so that the nodes have
> different channel offsets for the different slots within the schedule.

There can also be multiple networks (or parts of network) using
different set of slotframe lengths, and channel offsets etc, and if
things are set properly they do not collide.

I.e. if you have network(s) using same hopping sequence and hopping
sequence length, but different number of slots (slotframe length) and
different channel offsets, they can coexists in same space, and they
will not transmit on the same channel.

I made an excel sheet to demonstrate this:

https://mentor.ieee.org/802.15/dcn/15/15-15-0604-00-0mag-example-of-tsch-schedule.xls

Here you can have multiple networks (A-J) and they can have different
number number of slots and different channel offset and you can see
that do not collide if parameters are selected properly (i.e. they use
different channel offset).

Using different channel offsets you can configure network with
different slot frame lengths, i.e. you can have one with slotframe
length of 100 to allow lots of clients, and another one with slotframe
length of 5 to allow fast access to media etc.

If device is part of multiple of those it will pick which of those
networks it is using based on priority, i.e., if it wants to transmit
something and has slot that allows transmit on one network, it tunes
in to that channel and transmits. Otherwise it picks the one of the
receive channels and listents to that etc.

802.15.4 has some text about this, but mostly this is defined by the
vendor making the devices...

> > - ASN is used to construct a nonce when securing link-layer frames.
> > Security is such that we never want to re-use the same nonce.
> 
> Although you could construct the nonce to be (ASFN * nSlots +
> slotOffset).  And you need to do a construction like that to make sure
> that different nodes with different channel offsets during the same slot
> do not use the same nonce.

Nonce also has the transmitters extended address, which makes the
nonces different for each transmitter, and one transmitter can only
transmit one packet on each slot thus nonce is unique. 

> I suspect the main reason is that several slotframe schedules of
> different lengths can be active at the same time in one network.
> Keeping a separate ASFN for each schedule would require a lot of work.

Yes.
-- 
[e-mail]