[net2-wg] minutes from February 23 teleconference

[Omprakash Gnawali]

Present: Phil, Om, Rodrigo, Joe, Gil, Jonathan

Meeting started at 9:10am

Phil: I can work on writing up code in the 2nd half of March. Can
commit to finish dissemination by the end of March.

Rodrigo: We haven't started writing code either.

Phil: We have Trickle timer code. I can say I want a trickle timer,
instantitate on parameter. I can also say I want 7 of these timers.

Om: Is there a schedule for TinyOS 2?

Phil: I don't think there is schedule. A couple of groups in Europe
are working on different platforms. There is a list of things that
need to be done for it to be released. We also need a lot of feedback
from the community. Sometime during the summer is a reasonable
timeframe.

Om: Joe, is there any update on SP? We might be able to use some of
that for link estimation, neighbor management.

Joe: It will not be in TinyOS 1 or TinyOS 2 tree. It will be done next
week.

Rodrigo: Lets make the interfaces compatible with Joe, Moteiv SP.

Om: Where is the best place to look for SP interface?

Joe: Moteive code will be released next week. It will only work for
Tmote. Probably won't work with Xbow Telos or MicaZ. Release will be
on Moteiv website.

Om: Once it is released, I will look at the interfaces in more detail
and make the current neighbor and link estimation interfaces
compatible.

Rodrigo: Lets commit to end of March for collection.

Om: Sounds reasonable.

Rodrigo: There seem to be three proposals for longer term for this
working group. SP, specify a network architecture which will allow
many protocols, and Zigbee, which does not seem to raise a lot of
interest in this working group.

Phil: Zigbee 1.1 might be coming up according to Mulligan. That is a
reason to wait till 1.1 is out.

Joe: They won't move that fast. 1.0 effort will not be a waste but
some features might be optional in 1.1. For example there will be AODV
in Zigbee for sure.

Rodrigo: AODV will be a good first step then.

Joe: AODV will enable route discovery to type of data. Opposite of
Drip.

Phil: AODV is not a huge program once you have link estimation etc.

Om: IP is interesting for us? There seemed to be some interest from
what I heard during TTX.

Phil: Header compression can be done as Invensys has done and is
probably not challenging. Route discovery might be very challenging.

Rodrigo: IP will be point-to-point routing.

Phil: There is hierarchy in Zigbee so natural for hierarchical IP but
the name of the node might change.

Rodrigo: What does IP on sensornet mean? Send from the Internet to a
sensornet? Send from sensornet to the Internet? Both?

Phil: Sounds interesting but seems too open right now.

Om: I think we should focus on well understood problems in this
working group.

Phil: There are many implementations for SP on many platforms but not
on all platforms. It will be reasonable to set a goal to have an
overall SP for TinyOS. Also draw from experience to further refine the
interfaces and implementations.

Gil: I would like to see AODV and things like that because there are
enough people working on SP already.

Phil: Henri and others are working on radically different
radios. Congestion bit discussion. It will be interesting to combine
this knowledge and improve SP.

Jon: Will our work replicate other SP efforts?

Joe: I can commit to contributing SP provided Apache license gets
adopted.

Rodrigo: Berkeley SP will have BSD or something like that, Moteiv will
have Apache license. Is there a way to simplify that?

Joe: The implementations might not directly import but the interfaces
will. You guys can look at it and use it if you like.

Om: What if we like your code? Can we put that in the TinyOS 2 tree?

Phil: Probably not. BSD says include attribution in some end user
material. Apache says include attribution in user
documentation. Academic license does not require attribution in the
documentation and only requires author list in the source file. The
alliance is looking deeply into these issues.

Gil: I have dissemination implementation that compiles and runs and it
is being tested.

Meeting ended at 9:58am