Meeting Notes 08/08/2008
* CTP white bit * net2 plans
* Phil, Stanford * Mike, JHU * Razvan, JHU * Om, USC * Kaisen, UCSD * Romain, Amadeus, France
* white-bit testing (Om) * Need a web page that describes all the protocols we have * More extensive Deluge documentation (Razvan) * Tymo bug fixes (Romain)
Om: White bit not being used in the code in the CVS. I am testing after adding white bit.
Phil: Why is white bit not there?
Om: Probably got lost during code refactoring when we went from the version for the paper to the version for CVS.
Om: Lets discuss for each protocol, where we stand, and where we plan to go.
Om: Who is using it?
Phil: It has not been there for long enough. It is forward looking in the sense that it is designed for situation in which there are many items to dissemination so there might not be applications yet for DIP.
Om: Are we doing a good enough job advertising?
Om: We sent emails. Should we do other types of advertisement?
Phil: We should have a page that lists all the protocols we offer
Om: And a short description of each protocol before having to go into the details of the protocol.
Kaisen: One person has contacted me about DIP
Om: Are you open to making improvements?
Kaisen: Not sure. No planned improvements
Om: White bit experiment going on. What else do we need to do?
Phil: It needs to work on low power.
Om: Andreas is interested in a test suite.
Om: Some interface enhancements suggested for people who want more information about the routing state.
Om: Agree low power is the priority.
Phil: Once we have low power CTP that works well, people will start using the low power MAC so we get the benefit of testing the low power MAC more rigorously.
Phil: It has been stable for a year now.
Om: Are you interested in improving it?
Phil: Definitely not. No planned improvement but open to fixing bugs.
Phil: Its status is same as that of MultihopLQI.
Om: How about dissemination of middle sized objects?
Kaisen: What is middle sized?
Om: A few pkts so you don't want the application to keep track of state per pkt but don't want to use Deluge either. A Tenet task would be an example.
Phil: Virtual programs.
Kaisen: Could you not have applications manage state?
Om: You could but it would be nice if you didn't have to just like you don't have to if you disseminate a small object. You think of it has an object and you disseminate.
Phil: Does Deluge use Drip for commands as well as metadata?
Razvan: It has its own timers for the blocks.
Razvan: Here are the things I have in mind for Deluge: document reprogramming part, tosboot still looks like a TinyOS 1.x program, volume manager could be a bit more generic, support one hop reprogramming, Typhoon related features such as channel switching
Om: Channel switching might interact with many other things running on the mote so it might not be straightforward.
Phil: Could be dangerous without a mechanism to coordinate between different components.
Razvan: It already works but we haven't looked at imote2 yet.
Om: Do you have imote2's?
Om: My guess based on what I read, there is a high demand for Tymo. Generally, people come forward and say they want their protocol in. This might be a case when we should really bring Tymo up to release standard.
Romain: I am still interested in support and improving Tymo. I am answering questions that people ask privately so it might seem like I have disappeared but I am still in.
Om: Phil had some feedback about some bugs in Tymo
Phil: Brano had to fix some problems so it would be good to get those changes
Om: Anything else we want net2 to do in the next six months?
Razvan: Koala works great and will put it in contrib soon.