Using Hardware in the loop for real protocol testing

[Yagmur Konuslu]

Dear All,

i have real applications which will communicate using a custom protocol. The number of application instances is unknown or will change while running the simulation

(you may guess, application instances will be deployed on wireless nodes). Hardware in the loop seems like the right solution for connecting external apps to the simulator. If understood correctly the provided socket samples are showing the case for a single external application instance. For connecting multiple instances i guess i have to create a proxy module (module which will the gateway between the external app and the simulator) dynamically (since the number of external instances in unknown). My solution is as follows:

- RtSocketScheduler listens for connection requests

- Application connects

- RtSocketScheduler accepts the new connection, creates a proxy module, associates the module with the new socket, makes the connections between the proxy modules

- When application sends data, scheduler forwards the data to the module which further forwards the message to the simulation nodes.

- When a simulation nodes receives data from another one, it sends the data over the socket to the real external application.

My question is: Is this the way i should follow? Or do i have to dive deep into the simulator code and somehow try to merge my application code?

Thanks in advance

  

[Rudolf Hornig]

This seems to be feasible. For each connection request you create the corresponding modules in the simulator. (the nclients example works similarly (but it is purely inside the simulator).