Net2WG/Notes/20091204
Contents
Net2 Meeting Notes for 12/4/2009
Agenda
* Blip platform driver * DHV code / AM ID * Link Estimation
Participants
* Razvan, JHU * Mike, JHU * Phil, Stanford * Tao, UC Merced * Steve, UC Berkeley * Om, Stanford * Thanh, Portland State * Martin, Crossbow
Action Items
* Writeup for Blip project description (Steve) * Link estimator bug fix (Om) * DHV AM ID request (Thanh)
Discussion Notes
Blip
Steve - Lets not wait for the Cygwin Blip platform code for this release
Om - Is it worth doing it for you long term?
Steve - It is better if we can get somebody to work on it.
Om - Send a description of the project, I can forward the description to people looking for projects.
DHV
Thanh - Just need one AM ID.
Phil - DHV code handles the booted event and starts the radio. The model is applications should start or stop the radio, not the protocols.
Thanh - Was doing automatic radio on, but we can have the application do that.
Phil - Not necessary to make this change for this release.
Om - We should start thinking about using a single Trickle timer implementation across the dissemination protocols.
Link Estimation
Om - Tao wants to share his experience designing and implementing various link estimators.
Tao - Used le/4bitle/rnp estimators with ctp. Was able to easily implement rnp in this framework. White bit was not used in earlier version but seems to be used now.
Phil - White bit is provides additional information about a link. This information can be used to tag a link as usable even though we might not have received many packets on that link.
Tao - Compare bit interface was confusing and had to send emails to Om about that. This bit seems quite useful.
Tao - RNP and ETX are two different metrics so it is not possible to mix and match, however both are additive so it is possible to reuse the routing engine. But the thresholds and constants might need adjustments depending on the metric used.
Misc Notes
Om - will check in the bug fix to the link estimator. The bug fix allows the nodes to send packets to the root even when the root is plugged in and out. Bug found by Maria.
