Meeting Notes for 4/16/2010
* CTP experiments and deployments * Future projects
* Rodrigo, Brown * Mike, JHU * Om, Stanford
* Setup CTP page * Think about CTP workshop
Rodrigo: We did some CTP experiments with Quanto. CTP creates hotspots, particularly because nodes with poor connectivity try too hard to reach their parents.
Rodrigo: Number of tx is a poor proxy for energy consumption on a node that just runs CTP, but duty-cycle is a good proxy for energy consumption for such applications.
Om: We should make CTP more energy aware.
Rodrigo: Lets work on the design and think about implementation later.
Om: We ran CTP for 3 months. One interesting observation is the diurnal pattern - churn and cost are high during the day and low at night. The pattern repeats. Weekends have low churn and cost as well. We could have guessed but had not observed this pattern before.
Rodrigo: Considering the efforts at ROLL, what is CTP's future?
Om: Unencumbered research vehicle, robust open source alternative, and that it informed some design decisions in ROLL.
Rodrigo: We should create a page that talks about CTP, related work, and open questions.
Om: I will start on that and send an email.
Mike: Are we still working on loops?
Rodrigo: We are not currently working on loops. It is a project we should keep in mind.
Rodrigo: About the workshop, what is the idea? Get-together of protocol implementors and CTP users?
Om: We can broaden it a bit to collection protocols (eg., Koala type protocols) but restrict it to experiences and insights from implementations and deployments.
Rodrigo: Lets writeup a CFP and send it to a few people to gauge the interest.
Mike: I will start thinking about framework for source routing.