We are running dataplane tests using netperf tool to measure latency during pod to pod communication. We observed latency of 100 usec with 4.13 OVN vs 190 usec in 4.14 OVN-IC. Client and server pods are hosted in different host in same AWS zone.
We can find more results here https://docs.google.com/spreadsheets/d/1TEx29Vn2L20bYp7sXgW1tVA7NpK16rKxLOjpnBLr0qo/edit?usp=sharing

OCP 4.14 4.14.0-0.nightly-2023-10-06-234925
OCP 4.13 4.13.0-0.nightly-2023-10-08-090900

Always

1. Deploy self-managed aws cluster (2 worker nodes sufficient)
2. run netperf using perf team's e2e script
a) git clone https://github.com/cloud-bulldozer/e2e-benchmarking
b) cd e2e-benchmarking/workloads/network-perf-v2/
c) workload=full-run.yaml ./run.sh
 

Double the latency seen with OVN-IC 

10% deviation in latency is acceptable

Same configuration is used in both the environments i.e
1. worker count - 24 (2 workers sufficient for this testing. This test makes sure that pods are scheduled on different worker nodes before it runs uperf tool)
2. master and worker instance types - m5.2xlarge