Better performance with HTTP/2
summary
HTTP/2 replaces many parallel TCP connections with multiplexing over one TCP connection, cutting connection setup overhead and TCP slowstart. Benchmarks published on GitHub show dramatic capacity gains: single-connection servers sustain thousands of concurrent streams where old 6‑connection limits exhausted servers and caused connection churn. A Vert.x HTTP/2 server implementation uses the event‑driven Vert.x model and multi‑core deployment to turn latency into throughput, lowering median response time in benchmarks and raising requests‑per‑second substantially versus HTTP/1.1 setups. Practical code and results are available for direct replication.
AI-written summary · about
chapters (13)
- 0:00HTTP/2 introduction
- 3:58HTTP/1.1 performance issues
- 7:19Why HTTP/2 was created
- 9:27Multiplexing in HTTP/2
- 13:50Compression benefits in HTTP/2
- 17:12Response time and backend impact
- 20:07Capacity planning and concurrency
- 24:51Polyglot frameworks and reactivity
- 29:18Request handling in HTTP/2
- 35:00Async callbacks optimization
- 40:47Handling many connections
- 44:00HTTP/2 tools and ecosystems
- 47:13Conclusion
