fasthttpGo 的快速 HTTP 包
fasthttp 是 Go 的快速 HTTP 实现,当前在 1M 并发的生产环境使用非常成功,可以从单个服务器进行 100K qps 的持续连接。
HTTP 服务器性能与 net/http 比较
总而言之,fasthttp 比 net/http 快 10 倍
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
评论
clj-httpClojure的HTTP客户端包
clj-http是一个新的ClojureHTTP客户端开发包,主要特点是设计简单、可靠、可扩展和可测试。示例代码:(require'[clj-http.client:asclient])(client
clj-httpClojure的HTTP客户端包
0
moa-router一个快速的 HTTP 路由器
一个快速的HTTP路由器,在内部使用一个高性能的基数树(RadixTree )(又名紧凑前缀树(PrefixTree)),支持路由参数,通配符,它是独立的框架。性能$autocannon127.0.0
moa-router一个快速的 HTTP 路由器
0
FleaHttpd小型快速的HTTP服务器
fleahttpd:“跳蚤服务器”,是一个用C从头写的一个轻量级网络伺服器,静态文件的抓取速度大概是Apache的三倍。查看FleaHttpd与Apache之间的性能比较结果。
FleaHttpd小型快速的HTTP服务器
0
rails-asyncweb快速 Http 服务器
rails-asyncweb 是用于 JRuby on Rails 的一个快速 Http 服务器。
rails-asyncweb快速 Http 服务器
0