Hi Romain!I've written new test program. It uses either ActiveMessage or DymoNetwork to transfer messages via radio.
The used ifdefs are as small as possible to have a good (I think) compare. On this little program can be already seen a different behavior of two protocols.
Only one thing I encounter only on motes (but also with this small application). I could figured out, that after delay 40 s the message will be sent, sendDone() says FAILED, but the message will be received from remote mote, sent back and successfully received by "basestation" mote. If I send message again, I see the same scenario.
When I for the first time send a messages, it gets about 120 ms to return. Than only 20...60 ms. But after this delay all messages return after about 120 ms. The same time needed when there is no known routes, it seems when the routes are deleted and needed once more it doesn't work as expected. Or is this something else?
In sources I attach, there are some more comments about problems. With make -f Makefile.Driver make -f Makefile.Driver allI generate TestTymoLoop.java the client program and TestTymoLoop C++ TOSSIM-SF program for simulation. In Makefile with commenting (out) MULTIHOP you can switch between ActiveMessage and DymoNetwork.
The client program I start with java TestTymoLoop -comm [EMAIL PROTECTED]:9001 Could you please suggest me something? Best wishes, Andrey Andrey Gursky wrote:
Hi Romain! Here are details: In Makefile I've commented: #CFLAGS += -DCC2420_NO_ACKNOWLEDGEMENTS Otherwise the things are much worse.The application is derived from BaseStation, changed so that the once sent message (received throw serial from PC) arrives the TARGET comes back TO SOURCE and is transmitted throw serial back to PC.Soon I will extend the TestTymoAG I've sent to you couple of weeks ago to achieve the minimal example that stops to work as expected.The results of tests with 2 motes: After I send a message I'm waiting ... seconds than transmit a new message If less than 30 s: --> OK ---------- approximately 35 s:it seems that the message I once transmit throw serial had being sent, but there was not acknowledgement so the SOURCE transmits it again, then receives it back, but the was still no acknowledgment for successfully sending, so it repeats it again and again.after 450ms received packet sequence number 78 after 658ms received packet sequence number 78 after 866ms received packet sequence number 78 after 1059ms received packet sequence number 78 after 1266ms received packet sequence number 78 after 1458ms received packet sequence number 78 after 1874ms received packet sequence number 78 after 2274ms received packet sequence number 78 after 2370ms received packet sequence number 78 after 2578ms received packet sequence number 78 after 2786ms received packet sequence number 78 after 2993ms received packet sequence number 78 ------------------- more than approximately 40 sIt gets worse. The message throw serial comes also many times but now defect.receive error for TestTymoLoopMsg (AM type 9): invalid length message received (too short) receive error for TestTymoLoopMsg (AM type 9): invalid length message received (too short) receive error for TestTymoLoopMsg (AM type 9): invalid length message received (too short) receive error for TestTymoLoopMsg (AM type 9): invalid length message received (too short) receive error for TestTymoLoopMsg (AM type 9): invalid length message received (too short)----------So the same application works only if the time between transmission is less than 30 s. So I thought that the problem could be within deletion of old routes. Or you could see already another reasons?If I use 3 motes after some time I make no transmissions there is only one message that comes back defect. After that is all OK till the next long pause between transmissions.Best wishes, Andrey Andrey Gursky wrote:Hi Romain and Varun!I can confirm that (Test)Tymo works (at least) with 2 motes and 3 motes (multihop).But there are other surprises, I'll describe detailed later. Briefly there seems to be a problem with removing after timeout old routes. The bad thing: this occurs only on motes not in TOSSIM. So I try to decide what data I could and should collect to post here that can help to investigate this issue.Regards, Andrey Romain Thouvenin wrote:On Mon, Jul 14, 2008 at 4:22 AM, Varun Jain <[EMAIL PROTECTED]> wrote:Hi, I am trying to run the TestTymo application. I have programmed two nodes with nodeId 1 and 3. I have a Push button on my motes so I have modified the application such that I send a packet from the 'Origin' only when an event (in this case Pushbutton) is generated. On the 'target' end, I make an LED toggle when the packet is received correctly and it works but I want to check Multihop behavior of Tymo now. So I programmed another mote with nodeID = 7. I kept the 'target node (3) ' out of range of the ORIGIN (1). Now, I kept node-7 such that it is in range of other two motes. But there is no multihop communication taking place whereas due to the Intercept interface defined in "TestTymo" application, it should hop from 1 to 7(which it should forward) and then to 3 but it is not. Has someone encountered with this problem??Hello, Do you toggle a LED on the intermediate node to check that it receives/forwards something? Anyway, I just tried in TOSSIM to set a 4-node network like this one: 1 -- 3 | | 4 -- 2 and it worked fine. Unfortunately I don't have motes to compare with the real-world behaviour. Romain _______________________________________________ Tinyos-help mailing list [email protected] https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help_______________________________________________ Tinyos-help mailing list [email protected] https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help_______________________________________________ Tinyos-help mailing list [email protected] https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
testTymoLoop.tar.gz
Description: GNU Zip compressed data
_______________________________________________ Tinyos-help mailing list [email protected] https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
