Fast HTTP implementation for Go.
Currently fasthttp is successfully used in a production serving up to 1M concurrent keep-alive connections doing 100K qps from a single server.
HTTP server performance comparison with net/http
In short, fasthttp is up to 10 times faster than net/http. Below are benchmark results.
GOMAXPROCS=1
net/http:
$ GOMAXPROCS=1 go test -bench=NetHTTPServerGet -benchmem
PASS
BenchmarkNetHTTPServerGet1ReqPerConn 100000 21211 ns/op 2407 B/op 30 allocs/op
BenchmarkNetHTTPServerGet2ReqPerConn 100000 15682 ns/op 2373 B/op 24 allocs/op
BenchmarkNetHTTPServerGet10ReqPerConn 200000 9957 ns/op 2103 B/op 19 allocs/op
BenchmarkNetHTTPServerGet10000ReqPerConn 200000 8243 ns/op 2034 B/op 18 allocs/op
BenchmarkNetHTTPServerGet1ReqPerConn1KClients 50000 23474 ns/op 2704 B/op 30 allocs/op
BenchmarkNetHTTPServerGet2ReqPerConn1KClients 100000 18124 ns/op 2539 B/op 24 allocs/op
BenchmarkNetHTTPServerGet10ReqPerConn1KClients 100000 11815 ns/op 2689 B/op 19 allocs/op
BenchmarkNetHTTPServerGet10KReqPerConn1KClients 200000 9106 ns/op 2034 B/op 18 allocs/op
fasthttp:
$ GOMAXPROCS=1 go test -bench=kServerGet -benchmem
PASS
BenchmarkServerGet1ReqPerConn 500000 2495 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet2ReqPerConn 1000000 1925 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet10ReqPerConn 1000000 1300 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet10KReqPerConn 1000000 1140 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet1ReqPerConn1KClients 500000 2460 ns/op 1 B/op 0 allocs/op
BenchmarkServerGet2ReqPerConn1KClients 1000000 1962 ns/op 1 B/op 0 allocs/op
BenchmarkServerGet10ReqPerConn1KClients 1000000 1340 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet10KReqPerConn1KClients 1000000 1180 ns/op 0 B/op 0 allocs/op
GOMAXPROCS=4
net/http:
$ GOMAXPROCS=4 go test -bench=NetHTTPServerGet -benchmem
PASS
BenchmarkNetHTTPServerGet1ReqPerConn-4 200000 5929 ns/op 2434 B/op 30 allocs/op
BenchmarkNetHTTPServerGet2ReqPerConn-4 300000 4153 ns/op 2399 B/op 24 allocs/op
BenchmarkNetHTTPServerGet10ReqPerConn-4 500000 2751 ns/op 2118 B/op 19 allocs/op
BenchmarkNetHTTPServerGet10000ReqPerConn-4 500000 2398 ns/op 2037 B/op 18 allocs/op
BenchmarkNetHTTPServerGet1ReqPerConn1KClients-4 200000 5979 ns/op 2494 B/op 30 allocs/op
BenchmarkNetHTTPServerGet2ReqPerConn1KClients-4 300000 4582 ns/op 2457 B/op 24 allocs/op
BenchmarkNetHTTPServerGet10ReqPerConn1KClients-4 300000 3589 ns/op 2537 B/op 19 allocs/op
BenchmarkNetHTTPServerGet10KReqPerConn1KClients-4 500000 2465 ns/op 2036 B/op 18 allocs/op
fasthttp:
$ GOMAXPROCS=4 go test -bench=kServerGet -benchmem
PASS
BenchmarkServerGet1ReqPerConn-4 2000000 1094 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet2ReqPerConn-4 2000000 707 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet10ReqPerConn-4 3000000 417 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet10KReqPerConn-4 5000000 351 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet1ReqPerConn1KClients-4 2000000 916 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet2ReqPerConn1KClients-4 2000000 655 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet10ReqPerConn1KClients-4 3000000 404 ns/op 0 B/op 0 allocs/op
BenchmarkServerGet10KReqPerConn1KClients-4 5000000 359 ns/op 0 B/op 0 allocs/op
- Use reuseport listener.
- Run a separate server instance per CPU core with GOMAXPROCS=1.
- Pin each server instance to a separate CPU core using taskset.
- Ensure the interrupts of multiqueue network card are evenly distributed between CPU cores. See this article for details.
- Do not allocate objects and buffers - just reuse them as much as possible. Fasthttp API design encourages this.
- sync.Pool is your best friend.
- Profile your program
in production.
go tool pprof --alloc_objects your-program mem.pprofusually gives better insights for optimization thango tool pprof your-program cpu.pprof. - Write tests and benchmarks for hot paths.
- Avoid conversion between []byte and string, since this may result in memory allocation+copy. Fasthttp API provides functions for both []byte and string - use these functions instead of converting manually between []byte and string.