[Tinyos-devel] two zigbee implementations

[David E. Culler]

My theory is that if the working group has a focus and direction that 
people are enthusiastic about critical mass will form around it.  If it 
is buried in existing groups that already have a lot to do, it will bump 
along sub-critical.  The primary determiner is some people to step 
forward to chair or co-chair.

Assuming that happens, there will be some interesting points of 
intersection.  Specifically,

1. If an 15.4 LP MAC group forms and produces a high quality MAC it 
would be natural for NET to put its protocols on top of it.  Thus, it 
would make a lot of sense to get buy in on the proposed MAC design 
before a huge amount of work goes into implementing it.  If, in 
addition, a Zigbee group forms and if that group is implementing the 
Zigbee mesh protocol, it would have a similar interest.

2. If a Zigbee group forms, it will need to make some decisions about 
how it utilizes the work of other working groups.  Does it utilize NET 
protocols or implement an additional one.  If a new mesh protocol is 
done, perhaps that would best be done jointly with NET, so it can 
leverage a lot of the same substructure.

Even with joint activity on the mesh protocol, there is plenty of other 
stuff for the zigbee working group, as most of the implementation goes 
into the ZAPS and ZCL upper layers.  Indeed, these layers could be 
hosted over a variety of transports. 

Most of these issues would naturally be clarified in the working group 
charter and membership documents - assuming folks step forward to kick 
off the groups.

D.

Mário Alves wrote:
>
>
> Philip Levis wrote:
>
>>
>> On Jul 11, 2008, at 9:44 AM, David E. Culler wrote:
>>
>>> 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.
>>>
>>
>> I agree in theory -- the only drawback to this breakdown is whether  
>> there's enough critical mass for working groups to form around the  
>> topics. This historically is why tasks like 15.4 and zigbee have  
>> fallen under existing WGs (core and net2, respectively); doing so 
>> gave  the developer (each had a single developer) a group of other 
>> folks  working on somewhat similar problems with whom they could talk 
>> and  collaborate. That being said, now that WGs have moved their  
>> discussions to -devel, this might be much less of an issue than it  
>> once was.
>>
>> Phil
>>
> Honestly, I see relevant points in both David's and Philip's comments.
>
> Note that our (Porto) work goes far beyond the default IEEE 
> 802.15.4/ZigBee protocols as defined in their standard specifications. 
> We have been proposing, implementing and validating new mechanisms to 
> enhance their behaviour and that are backward compatible with their 
> standard specifications.  So, not only the "default" protocols will be 
> available within the TinyOS community, but also these add-ons.
>
> Surely there is no sense of forming WGs with only one 
> contributor/institution each. Thus, if there are neither enough 
> momentum nor critical mass (at least 3-4 contributors/institutions), 
> there is no motivation for the (new) 15.4 and ZigBee WGs to see the 
> light. Even if such a critical mass exists, I wonder if a separate 
> ZigBee Working Group should be formed. Then why not creating separate 
> WGs dedicated to ISA SP100, WirelessHART or 6LoWPAN? Moreover, looking 
> at the other existing TinyOS WGs, they all (unless eventually 8051) 
> seem quite broad (core, net2, sim, testbed,...). Thus, if keeping 
> ZigBee within the Net2 WG is not an option, I would go for just one 
> new 15.4/ZigBee WG (Option 3 of Omprakash).
>
> Ultimately, the final decision should be taken by TinyOS management 
> people, who have a more integrative view of what is being done by what 
> WG and also about future directions that should be followed.
>
> Cheers,
>
>    Mário
>
>