Net2WG/Notes/20100319
Contents
Net2 Meeting Notes for 3/19/2010
Agenda
* Future projects * CTP/4bitle bugs
Participants
* Phil, Stanford * Thanh, PSU * Mike, JHU * Om, Stanford
Future projects
Om - Some outstanding projects: BLS, Tymo. Then there is code cleanup.
Phil - When we started net2, we identified collection to be a basic building block. And we worked on it and we have a reasonable protocol. Are there other basic protocols?
Om - I would put Dissemination in that category. We have three options and the software is stable.
Phil - Maybe reliable transport is the next step?
Mike - Koala?
Phil - Source routing can be useful for a lot of applications and protocols.
Mike - Koala already users source route.
Om - JHU has the most experience among people here at this meeting doing source routing. Koala is definitely a good case study on how to do source routing that works. Mike, are you interested in leading this source route effort?
Mike - Sure.
Om - Lot of scenario seem to want command-response type of pattern. Most likely this is a composition of dissemination and collection but it will be nice to have an application that demonstrates how to do this.
Om - I hear a lot of demand for clustering protocols. It is unclear what it means though. Not sure if we can generalize it enough to make it useful for a lot of scenarios.
Thanh - It is not even clear if it is a networking protocol.
Om - RPL?
Phil - JeongGil is working on RPL implementation. It will be great to include it in net2.
Om - Definitely more relevant for net2 than even 6lowpan.
Thanh - I am working on reasoning about latency in a network and protocols. I will also commit some enhancements to DHV.
Om - We would love to hear about your work on latency when you are ready.
Om - We should also be open to other collection, dissemination, etc. protocols.
Phil - Definitely.
Om - The thresholds for software quality will still be the same.
CTP/4bitle bugs
Om - Summary of bugs and deadcode reported in the last few weeks: buggy queue check in the forwarding engine, unnecessary ECN bit set while snooping a packet, dead code in link estimator where it checks for unrealizable packet gap, redundant assignment of last sequence number from a given source. Is it ok to check them in?
Phil - Yes, the release has already been tagged.
Misc
Om - Should we switch to every other week meeting by default?
Phil - Lets stick with every week.
