Pipy proxy helps improve SpringBoot REST Service QPS & Latency
Spring and SpringBoot development platforms are one of the most popular development platform among Java developers.
We performed a simple benchmark of SpringBoot starter application running with default configurations with and without a sidecar Pipy proxy and measured QPS and Latency indicators.
Our results shows that deploying Pipy proxy as sidecar improved QPS and reduced latency by almost 10% in our testing environment.
Test Results
| # | OS | Tool | Mode | QPS | Latency P50 | P90 | P99 | P99.9 |
|---|---|---|---|---|---|---|---|---|
| 1 | ubuntu | ab | springboot | 46102 | ||||
| 2 | ubuntu | ab | pipy-->springboot | 50303 | ||||
| 3 | ubuntu | wrk | springboot | 71840 | 1.30ms | 2.04ms | 4.27ms | |
| 4 | ubuntu | wrk | pipy-->springboot | 73055 | 1.35ms | 2.09ms | 2.49ms | |
| 5 | ubuntu | fortio | springboot | 62249 | 0.0015217 | 0.00293947 | 0.00428543 | 0.00777612 |
| 6 | ubuntu | fortio | pipy-->springboot | 69233 | 0.0014350 | 0.00249741 | 0.00296226 | 0.00362949 |
- From the test results, we observed: Accessing Springboot service via Pipy proxy results in higher QPS and lower latency as compared to directly accessing the SpringBoot service.
For detailed test reports along with benchmark procedure, software used, load testing tools, and scripts used refer to Pipy wiki link below: