On GENI, when deploying the topology described in the coursera assignment I found some weird latency. It turns out that all of the packets were being sent to the controller repeatedly. What means that the controller is not successfully installing the flows.
Maybe we should do some unit tests to figure if the HP switch supports the flow mods used by SDX.
On GENI, when deploying the topology described in the coursera assignment I found some weird latency. It turns out that all of the packets were being sent to the controller repeatedly. What means that the controller is not successfully installing the flows.
Maybe we should do some unit tests to figure if the HP switch supports the flow mods used by SDX.
Here is a packet capture of the same experiment working correctly: https://github.com/sdn-ixp/sdx/blob/FlavioGeni/examples/GeniIXP/correct.pcapng.gz
Here is a packet capture of the packets repeatedly arriving at the controller: https://github.com/sdn-ixp/sdx/blob/FlavioGeni/examples/GeniIXP/flow.pcapng.gz?raw=true