Many-to-one & source routing for large scale networks:
- the concentrator sends the many-to-one route discovery --> all the nodes learn the route to the concentrator, but the concentrator doesn't know the routes to them
- for a high ram concentrator, the nodes will send a route record command before the first data packet
- when this route record command is received by the concentrator, the high ram concentrator will save the route and used for packets to that node --> after the route record was sent, the node can communicate with a high ram concentrator without the need of a route record command
- when the concentrator sends another many-to-one route request, all the nodes in the network will refresh the route table entry and send the route record command again before the first data packet
- to avoid sending the route record command multiple times the many-to-one route discovery should be done as rarely as possible
- however, if the network topology changes and the many-to-one route request is not sent, the data send using source routing will fail to reach the destination and if enough packets fail unicast route discovery would be initiated
Configurations associated to this topic:
•gConcentratorFlag_d
•gNwkHighRamConcentrator_d
•gNwkInfobaseMaxSourceRouteTableEntry_c
•gTotalSmallMsgs_d
•gConcentratorDiscoveryTime_c
•gNwkInfobaseMaxRouteTableEntry_c
