Cloud Performance Tests - Latency and Throughput

The goal of this project is to benchmark and compare the performances of load balancers in the areas sdn1 and octavia_t, based on latency and throughput. Results were obtained for the cases of 5 Clients with N Servers and N Clients with 50 Servers, with N being 1, 5, 10, 25 and 50. From these results, it seems that: Latency (the average) seems to be similar in both regions; Throughput (the average) seems to be smaller in the octavia_t region; 25.000 was the observed limit for Requests per Second. The conclusions are then that the tests could be further validated by rerunning them, as well as, going forward, that the team should be mindful of the limitations found and possibly investigate where the performance issues come from if they start having an impact on real use case