You have the story basically right -- when a new mote comes up, it solicits rapidly to find a network and join. Once it joins, it transmits link information to the LBR to allow it to form a topology. Once the node as formed stable links, it sends topology updates at a reduced rate. Thus, if the LBR reboots, it will have to wait for all the motes to report in which can be a while.
There are two basic options for getting around this. The first is to "rebuild" the DAG by incrementing the DAG sequence number (type "rebuild" in the router console). When a mote receives a router advertisement with a higher sequence number, it throws out all its link state and reconstructs everything (like it just booted). This is pretty heavyweight as it causes a global recomputation. Alternatively, you can change TGEN_MAX_INTERVAL in IPDispatch.h. This is the maximum interval between topology updates sent by motes. It's currently set to 5 minutes to keep the overhead low. The topology updates are opportunistically piggybacked on application data, so if your application sends data more frequently, you can decrease this without much added overhead (as long as you don't cause fragmentation). Steve On Mon, Jun 21, 2010 at 3:48 AM, sripada kadambar <sri.kadam...@gmail.com> wrote: > Dear All, > We are using the BLIP library for implementing IPv6 communication with > motes. The blip applications work pretty fine except for one problem; The > network formation takes a long time in case if the basestation undergoes a > reboot. In all the other cases such as node reboot no delay is encountered > and the node associate with the basestation quickly. So i was going through > the blip library trying to fix the problem. > initially when the network is powered up the network formation is quick as > all the nodes are transmitting solicitation messages. but once a node > associates with the basestation, it seems that nodes stop sending > solicitation messages altogether. So in case if the basestation undergoes a > reboot, it fails to detect the motes in the network as it no more recieves > solicitation messages from the motes. But ultimately the nodes are > recognized, though it takes a very long time to happen so. > Please let me know if my understanding of the code is correct, and if there > is any way to fix the problem. > Regards, > Sripada Kadambar > _______________________________________________ > Tinyos-help mailing list > Tinyos-help@millennium.berkeley.edu > https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help > -- stephen dawson-haggerty http://cs.berkeley.edu/~stevedh uc berkeley wireless and embedded systems lab berkeley, ca 94720 _______________________________________________ Tinyos-help mailing list Tinyos-help@millennium.berkeley.edu https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
