最近由于产品业务出现请求瓶颈,需要更换产品框架，针对现在集中主流方案进行了逐一测试  

服务器硬件配置:2 核 2G虚拟机 , 10000请求 500并发测试结果如下

性能测试结果

1、Nodejs Express测试结果如下 大约每秒处理2100请求 响应时间在230ms左右，cpu负载百分之40左右

Server Software:
Server Hostname:        172.16.200.202
Server Port:            3000

Document Path:          /
Document Length:        12 bytes

Concurrency Level:      500
Time taken for tests:   4.692 seconds
Complete requests:      10000
Failed requests:        0
Total transferred:      2060000 bytes
HTML transferred:       120000 bytes
Requests per second:    2131.47 [#/sec] (mean)
Time per request:       234.580 [ms] (mean)
Time per request:       0.469 [ms] (mean, across all concurrent requests)
Transfer rate:          428.79 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        1   16   8.3     15      57
Processing:    71  212  31.0    204     359
Waiting:       71  211  31.0    203     359
Total:         86  228  29.0    221     380

Percentage of the requests served within a certain time (ms)
  50%    221
  66%    230
  75%    241
  80%    248
  90%    258
  95%    291
  98%    302
  99%    304
 100%    380 (longest request)

2、PHP7+Swoole 秒处理请求2300  响应时间40s左右 cpu只占用了15% 

Server Software:        12XueWeibo
Server Hostname:        172.16.200.202
Server Port:            3456

Document Path:          /
Document Length:        27 bytes

Concurrency Level:      100
Time taken for tests:   3.406 seconds
Complete requests:      8000
Failed requests:        0
Total transferred:      1584000 bytes
HTML transferred:       216000 bytes
Requests per second:    2348.89 [#/sec] (mean)
Time per request:       42.573 [ms] (mean)
Time per request:       0.426 [ms] (mean, across all concurrent requests)
Transfer rate:          454.18 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        4   18   2.9     19      28
Processing:    10   24   3.5     24      46
Waiting:       10   24   3.5     24      46
Total:         16   42   3.4     42      60

Percentage of the requests served within a certain time (ms)
  50%     42
  66%     44
  75%     44
  80%     45
  90%     46
  95%     48
  98%     49
  99%     51
 100%     60 (longest request)

3、Nodejs Sails  每秒处理500左右请求 每个请求响应用掉190ms 时间 cpu 百分之50左右

Benchmarking 172.16.200.202 (be patient)
Completed 800 requests
Completed 1600 requests
Completed 2400 requests
Completed 3200 requests
Completed 4000 requests
Completed 4800 requests
Completed 5600 requests
Completed 6400 requests
Completed 7200 requests
Completed 8000 requests
Finished 8000 requests


Server Software:
Server Hostname:        172.16.200.202
Server Port:            1337

Document Path:          /
Document Length:        10444 bytes

Concurrency Level:      100
Time taken for tests:   14.820 seconds
Complete requests:      8000
Failed requests:        0
Total transferred:      87740708 bytes
HTML transferred:       83552000 bytes
Requests per second:    539.81 [#/sec] (mean)
Time per request:       185.249 [ms] (mean)
Time per request:       1.852 [ms] (mean, across all concurrent requests)
Transfer rate:          5781.69 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        2   12  14.3     11    1231
Processing:    23  173  48.9    165     515
Waiting:       16  125  40.9    123     432
Total:         34  185  50.2    176    1313

Percentage of the requests served within a certain time (ms)
  50%    176
  66%    179
  75%    181
  80%    184
  90%    239
  95%    298
  98%    343
  99%    362
 100%   1313 (longest request)

4、golang Beego   每秒大约处理2400个请求  每个请求响应时间45ms  cpu 百分之15左右

Document Path:          /routerinfo
Document Length:        1152 bytes

Concurrency Level:      100
Time taken for tests:   3.494 seconds
Complete requests:      8000
Failed requests:        7641
   (Connect: 0, Receive: 0, Length: 7641, Exceptions: 0)
Total transferred:      10348153 bytes
HTML transferred:       9468153 bytes
Requests per second:    2289.46 [#/sec] (mean)
Time per request:       43.678 [ms] (mean)
Time per request:       0.437 [ms] (mean, across all concurrent requests)
Transfer rate:          2892.06 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        2   17   4.5     18      27
Processing:    11   26   8.0     24      85
Waiting:       11   26   8.0     24      85
Total:         15   43   6.9     42      96

Percentage of the requests served within a certain time (ms)
  50%     42
  66%     44
  75%     46
  80%     46
  90%     49
  95%     54
  98%     64
  99%     71
 100%     96 (longest request)

5、 ThinkPHP 请求量缩减10倍1000, 并发量缩减50倍10,结果如下 每秒106请求 平均响应时间 94ms ,目前api就是用的此框架 果断需要更换了

Server Software:        Tengine/2.1.2
Server Hostname:        api.uia.com
Server Port:            80

Document Path:          /index/index/index
Document Length:        622 bytes

Concurrency Level:      10
Time taken for tests:   9.405 seconds
Complete requests:      1000
Failed requests:        0
Total transferred:      960000 bytes
HTML transferred:       622000 bytes
Requests per second:    106.33 [#/sec] (mean)
Time per request:       94.047 [ms] (mean)
Time per request:       9.405 [ms] (mean, across all concurrent requests)
Transfer rate:          99.68 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        1    2   0.6      2       8
Processing:    32   92  14.5     91     170
Waiting:       32   92  14.4     91     170
Total:         33   94  14.5     93     172

Percentage of the requests served within a certain time (ms)
  50%     93
  66%     99
  75%    103
  80%    105
  90%    111
  95%    117
  98%    126
  99%    133
 100%    172 (longest request)

很显然从结果看来 Sails、ThinkPHP等 经过框架路由之后 损耗还是很严重的！！