From TinyOS Wiki
Jump to: navigation, search


Net2WG/Notes/200802015 Meeting Notes


* TEP 118
* DIP Testing
* TTX projects


* Razvan, JHU
* Arsalan, Berkeley
* Om, USC
* Kaisen, UCSD
* Phil, Stanford

Action Items

* Send slides to Om

Discussion Notes

TEP 118

* 3 comment classes: interfaces, speed, and description

DIP testing

* Om: I did one three node test, and one thing was not clear to me. What is the difference between total items and new items?
* Kaisen: New items describes how many are new, total describes how many items there are. So new is a subset of total.
* Om: OK, it was not obvious to me. I was able to follow the directions and test on two nodes, and everything worked.
* Kaisen: OK, I also want to add a tool that will update it, like an injector. So the idea is that you have a base station that can send out an update. Rather than a set node.
* Om: What is DIPTestMaster?
* Kaisen: Well, I generate a source file from the description of the test you want to write.
* Om: I haven't tried it, but I should be able to switch the flags, and it should work for either DIP or Drip? 
* Kaisen: In theory, yes. Let me check and get back to you.
* Phil: Going forward, it can be nice to let people know about net2 by putting the path in the paper title.
* Om: Having a sequence number on the serial packets would be nice.
* Kaisen: Right -- there's the counter for how many items have updated.
* Om: OK, good, I tend to lose some serial messages on the testbed. Was this confusing just to me?
* Phil: I'm sure a sentence or two could clear it up.


* Om: Did people get my mail asking for slides?
* Razvan: Yes.
* Kaisen: Yes.
* Om: So let me talk about stuff we don't hear too much about. Zigbee, 6lowpan, and TYMO.
* Om: Let's start with TYMO. Romain has been sending me code and documents. I told him the key thing is having a test case or app. Romain wants to call in soon.
* Om: Zigbee and 6lowpan are both making progress.
* Om: So one other protocol: LQI. There's a bug in there that I found, where you turn off the radio and LQI locks. Phil said he'd fix it.
* Om: One last thing: Matt Welsh is willing to let us do long-term experiments on Motelab.
* Phil: That's great. One thing about Motelab. Nodes tend to come and go a little, so it's very hard to do repeated experiments. But for our purposes, we're testing, so this is not an issue.
* Kaisen: What about the serial interface? Is it 2.x compatible, or do you need the 1.x hack?
* Phil: In theory 2.x should work, but in practice I know Jung Il just uses the 1.x hack.
* Om: Razvan is also running CTP tests on his testbed.
* Razvan: Last time I ran, the maximum hopcount was 3. 
* Om: OK, send your slides, when you get a chance?