cacu http requests

[dos-group]

cacu http requests

[birnbaum]

Not sure if we had a reason for this, but why is the CACU implemented on top of SimPy? We could simply do a while True and send requests once per minute using sleep :)

[marvin-steinke]

Ah yes, no I copied that from your notebook. But still, if we want to achieve similiar results to the simulated scenario, I think the requests should be made as a thread to not fall behind the total runtime.

[birnbaum]

but the cacu as well as the API server are independent processes, they should have nothing to do with the simulation, no?

[marvin-steinke]

yes but we want to recreate the data from your paper. Now in your notebook there is no delay for setting the power mode. However, there is in the real cacu as of now. Should the cacu wait until the power mode is set in the VES or continue, i.e. the power mode request is executed as a thread? If the cacu is excuted on the same machine as the simulation this probably does not matter but I think we still need to cover other cases.

[marvin-steinke]

=> 60s for each "step". Determine a time delta for the step execution time and then time.sleep(60 - delta)