Limitation of OpenFlow pipeline - Packet Discards on Ingress Port on a 10G network

[Ramzah_Rehman]

Hello. I am using ARUBA 2930F switch. I have two DellPowerEdge 4700r servers connected via 10G SFP+ ports to the switch. I send TCP traffic at 10Gbps from one server to another. When I am not using OpenFlow, I get a throughput of ~9.4Gbps from one server to another. Also, I don’t see any Tx drops or Rx discards. However, when I use OpenFlow instead (enable OpenFlow standard pipeline), and send traffic again, I see a lot of Rx discards on the switch. As a result, the throughput fluctuates between 2Gbps to 8Gbps. Is it the issue with slower packet processing in the OpenFlow pipeline? I have made sure that the controller code and topology configuration is correct.

To sort of verify that the problem is with the OpenFlow pipeline, I enabled OpenFlow and sent UDP traffic at 2Gbps on the 10Gbps network and did not see any Rx discards.

I have been using the standard pipeline of OpenFlow. I have read in the documents about this behavior regarding the OpenFlow custom pipeline. It says that in OpenFlow custom pipeline: “packets are dropped at the ingress if the pipeline is busy processing too many packets. You can check Rx discards on the switch to obtain such data”. I believe a similar issue can arise in the standard pipeline as well.