Background (from BZ)

Instances on OVS DPDK currently cannot enable TSO (TCP segment offload), which greatly lowers their performance, making OVS-DPDK possibly a worse choice than not using it at all. Inter VM communication of iperf3 TCP throughput is very low on same host compare to non DPDK throughput. The root cause of the issue is that currently ovs-dpdk does not support TSO offload. With the offload enable, the throughput will be lower than traditional path.

The way we test with iperf3 is using TCP protocol, and currently DPDK doesn’t support TSO (TCP Segmentation offload) based on this link:
https://bugs.launchpad.net/devstack/+bug/1651727

In my opinion, we'd need to get the following into OVS-DPDK ASAP:
https://mail.openvswitch.org/pipermail/ovs-dev/2016-June/316414.html
https://ovs2016.sched.com/event/8aZf/optimizing-communications-grade-tcp-workloads-in-an-ovs-based-nfv-deployment-mark-kavanagh-intel
https://www.youtube.com/watch?v=hEP0_Bd3wrA

FID

https://docs.google.com/document/d/1o7e0HpTvEiz--7Tkpm3CdSLUdT3A1dGxCjOvvEXjYAY/edit?tab=t.0#heading=h.rv1bkehb3mtt

Usability

• The user can enable TSO on OVS-DPDK using EDPM
• The user can have better performance 

DOD

Development requirement:

• Enable and the feature in FDP with different nics
• Allow enabling TSO in openstack 

testing requirement:

• performance (iperf3/trex)
  • VM to VM on same host
  • VM to VM on different hosts
  • VM to VM connected with VLAN network
• NICs
  • Mellanox CX-5/6
  • Intel X710 - As of Nov 20th, Intel card are not supported if the traffic is tunneled with geneve/vxlan [ovs 3.3]
• functional testing
  • verify TSO is enabled in OVS