How can l handle the "NWK link status" message in the application layer?

Thanks for your answer Dirty Harry.

Yes you are right absolutely the "self-healing mesh network" is one of the most valuable things in Zigbee. But l got around to this point becauseof some problems l have experienced.  Actualy my main problem is not updated assoction list by the devices in their memories when occuring any changes in network. l wrote some codes and its requires restart the router when coordinator lost. Anyway  l decided to it is not suitable solution as you said.

When l switched off the coordinator in network includes router(s), it can not resume with same PANID due to the compiling the program without NV_RESTORE paremeter. But l can get a cleared assoction list. And when l compiled with NV_RESTORE the program, coordinator can carry on its mission with same PANID but old assocation list exist with no changes. l wrote some codes to update the asscotion list but sometimes its not work properly and caused to restart coordinator due to the invalid extended address input. Now l optimized my code as written below:

static void S5_update_assoc_list(void) {  

   uint8 qind = 0;  

   AddrMgrEntry_t addrEntry;   NLME_LeaveReq_t req;     while(AssociatedDevList[qind].shortAddr != 0xFFFF)  

  {    

    addrEntry.user  = ADDRMGR_USER_DEFAULT;    

    addrEntry.index = qind;     AddrMgrEntryGet( &addrEntry );        

    if(AssociatedDevList[qind].age > S5_DEVICE_AGE_LIMIT && AssociatedDevList[qind].shortAddr == addrEntry.nwkAddr)    

     {      

         req.extAddr = addrEntry.extAddr;      

req.removeChildren = TRUE;      

         req.rejoin = TRUE;      

         req.silent = FALSE;      

         NLME_LeaveReq( &req );    

      }    

    qind ++;  

    }

}

The code find devices with big age and remove from the association list. maybe it still needs to be optimized. For now it works fine.

Now l have a new problem... l setup a network includes one coordinator(C) and two routers(R1, R2) and R2 connected to R1, R1 connected to C (R2->R1->C) so Coordinator's association list consist only R1 address. When l switch of R1, l assume that R2 connected to C isn't it? but l can not see R2 address in coordinator list. Why?

ZC will be connected to PC via RS-232 in my project and l want to see updated Zigbee network topology on UART terminal. In "Method for Discovering Network Topology" PDF note the assocation table was used so l focused on it too. l am new in Zigbee and if there is another solution for discovering R2 as you said l will try. Thanks for your help.

Hi Harry again,

When l used "ZDP_MgmtLqiReq()" and "ZDO_ParseMgmtLqiRsp()" to get negihbor list, l can get only 3 number of neighbor device address eventough "neighborLqiEntries" indicates 6.

Why l can not get full list? For saving memory space?

Thanks

Hi,

I just stumbled on this thread, a bit late ;-)

I have a similar scenario, where maybe 200 ZEDs & ZRs will be sending application data to the ZC, which forwards the data to a remote server.

The easiest solution for having the network topology (map) is by each device sending its parent (next hop) as part of the application data.

NLME_GetRequest() works for ZEDs, but, as you explained - not for ZRs. If, while installing the devices, a ZR is moved, for any reason, is will still send it's original parent (which gave him his NwkAddr).

I've been going through the neighborTable, rtgTable and AssociatedDevList, but most of the time the data (neighborEntry_t.neighborExtAddr) is corrupted... ?

Any ideas?

See Router parent MAC - NLME_GetCoordExtAddr() for more details.

Thanks