[Tinyos-devel] two zigbee implementations

[David E. Culler]

If the community is willing to strive for defining a 15.4 mac that 
enables forwarding devices, not just RFDs, to operate at low power, I 
would not want to limit it to Zigbee.  Zigbee primarily has to do with 
applications profiles and a particular addressing and meshing.  It has 
little to do with the MAC per se.  In fact, it has largely forsaken 
utility in the MAC, setting for leaves turning themselves off whenever 
they feel like it and powered FFDs.

There are several organizations that have interest in MAC standard 
emerging - these include IEEE 802.15.4e, ISA sp100.11a, 6LoWPAN.  At 
some level, network layer interoperability is only relevant with link 
layer interoperability.  Right now we have several cases of potentially 
interoperable network, transport, and application layers, but they do 
slightly different things in the power management.

The question for a 15.4 working group is how much to take on.  There is 
no question that only supporting 16-bit short addresses is inadequate.  
EUID64 addressing is essential.  Presumably people are inspired to take 
on the power management aspect.  That doesn't necessarily mean getting 
sucked into the beacon morass.  Most groups have given up on this form 
of power management anyways.  The debate is primarily around (a) time 
scheduled communication vs sampled listening and (b) frequency agility.

Personally, I would like to see the TinyOS community grab ahold of the 
key challenges that so much of the rest of the world is grappling with 
and finding a hard time reaching a high quality solution, rather than 
slavishly implementing all the obscure features proposed in the IEEE 
802.15.4 spec which most of the world has already abandoned.  (I put PAN 
coordinators in this category too.)

Conversely, the ZCL cluster library is a very interesting application 
layer that does not need to be wedded to the Zigbee networking layer, 
much less the underlying.  It can be implemented over UDP or Active 
Messages. 

So, volunteers permitting, I would most like to see option 1.  Core 
should remain Core.  These other WGs should push Core to enable them.


Omprakash Gnawali wrote:
> Enumerating the options:
>
> 1. 15.4 WG. Zigbee WG.
> 2. 15.4 under Core. Zigbee WG.
> 3. 15.4/Zigbee WG
> 4. 15.4 under Core. 15.4/Zigbee under net2.
>
> Clearly 4 is the worst option which is why we started this discussion.
>
> I am not opposed to 3 because Zigbee will most likely be one of the
> most active user of 15.4 so it would be good for these two groups to
> work together tightly.
>
> The problem with 2 is Core might be too busy with developing multiple
> MACs - it already maintains the default MACs for each platform.
>
> I do not know if we have a critical number of people for 1 to be successful.
>
> So my vote is for 3 although 2 is also a good option.
>
> - om_p
>
> On Thu, Jul 10, 2008 at 7:31 AM, Mário Alves <mjf at isep.ipp.pt> wrote:
>   
>> Dear Omprakash, Vlado, Matt and David (et all),
>>
>> We went through the latest interactions and agree with your (Omprakash,
>> Vlado, Matt and  David) suggestions.
>>
>> So, our efforts can be reorganized in the following way:
>> - move all the IEEE 802.15.4 MAC implementation effort to the Core WG
>> - create the ZigBee WG to accommodate the ZigBee Network Layer
>> implementation (and remove this effort from the Net2 WG)
>>
>> We volunteer to drive the ZigBee WG and are willing to involve the following
>> people from our side:
>> - Shashi Prabh (Chair, in Cc to this e-mail)
>> - Ricardo Severino
>> - Nuno Pereira
>> - Mário Alves
>>
>> We expect participation from Berlin (Vlado and Jan: please confirm) in the
>> ZigBee WG, as well as any other parties/people that you might be aware could
>> contribute (e.g. ZigBee Security functionalities are not implemented yet),
>> including from your own groups.
>>
>> As the IEEE 802.15.4 MAC implementation effort is to be moved to the Core
>> WG, we are also available to commit to that effort/WG.
>>
>> Looking forward to read from you,
>>
>>   Mário
>>
>>
>>
>> David E. Culler wrote:
>>
>>     
>>> Omprakash Gnawali wrote:
>>>
>>>       
>>>> We have two zigbee implementations being developed by two groups -
>>>> Porto and Berlin. Would it make sense for these two efforts to either
>>>> form a new WG or put these two projects in the same wg (core or net2)
>>>> so that there is better interaction between the groups? Right now
>>>> Berlin works under Core while Porto works under net2 and this
>>>> arrangement seems an accident rather than planning.
>>>>
>>>> Thanks.
>>>>
>>>> - om_p
>>>> _______________________________________________
>>>> Tinyos-devel mailing list
>>>> Tinyos-devel at millennium.berkeley.edu
>>>> https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-devel
>>>>
>>>>         
>>> I think it would be good to have a zigbee working group.  Core and Net2
>>> already have so much on their agendas.  Do we have a volunteer to be chair?
>>>
>>>
>>>       
>> --
>> ==========================================
>> Mario Ferreira Alves
>> e-Mail: mjf at isep.ipp.pt
>> Polytechnic Institute of Porto (IPP)
>> School of Engineering (ISEP)
>> Rua Dr. Antonio Bernardino de Almeida, 431
>> 4200-072 Porto, Portugal
>> ------------------------------------------
>> IPP-HURRAY! Research Group/CISTER Research Unit
>> http://www.hurray.isep.ipp.pt
>> to me: /asp/show_people.asp?id=18
>> Tel: +351 22 8340502
>> Fax: +351 22 8340509
>> ------------------------------------------
>> Electrical Engineering Department
>> http://www.dee.isep.ipp.pt to me: /~malves
>> Tel: +351 22 8340500
>> Fax: +351 22 8321159
>> ==========================================
>>
>>
>>