Benchmark testing tool: ab

Benchmark testing tools (ab)

What is ab?
ab is a command line-based tool that can run on Windows and Linux platforms.
Why choose this tool?
Easy to use, just enter a few command options and get results, and easy to install.

Install ab

win: If there is an apache development environment locally, then ab is in the bin directory under the apache directory by default. You can see it by opening the bin directory
linux: Just yum install httpd-tools and install it

Easy to get started

Type the following code directly (note to switch the directory where ab is located)
-c 10 means the number of concurrency is 10
-n 10 for a total of 100 visits
Followed by the URL to be visited, remember to indispensable http://

D:phpStudyApachebin>ab -c 10 -n 1000 http://example.com/phpinfo.php

Wait for ab to run it yourself and get the following results

#ab tool Description, pay attention to the version number
This is ApacheBench, Version 2.3 <$Revision: 1604373 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licenseed to The Apache Software Foundation, http://www.apache.org/

#Test results, a total of 1000 tests, the access progress is displayed every 100 times, the website visited is example.com
Benchmarking example.com (be patient)
Completed 100 requests
Completed 200 requests
Completed 300 requests
Completed 400 requests
Completed 500 requests
Completed 600 requests
Completed 700 requests
Completed 800 requests
Completed 900 requests
Completed 1000 requests
Finished 1000 requests

#Server information
Server Software: nginx/1.6.2 #Server web software
Server Hostname: example.com #Host address
Server Port: 80 #Access port

#Document information
Document Path: /phpinfo.php #Accessed document
Document Length: 94804 bytes #Return the number of bytes of the result size (including the total number of bytes of all content in html, js, css, images and responses)

#Link information
Concurrency Level: 10 #Concurrency number, -c 10 sets 10 concurrencies
 Time taken for tests: 7.498 seconds #Total time taken, unit seconds
Complete requests: 1000 #The sum of the total number of completed requests in the total requests
Failed requests: 219 #The sum of the number of failed requests
 (Connect: 0, Receive: 0 , Length: 219, Exceptions: 0)
Non-2xx responses: 9 #The total number of successful requests in the 2xx series that were not received
Total transferred: 94135779 bytes #The total size bytes of the total response distance in the entire total number of requests, including header information
HTML transferred: 93952076 bytes #The total size bytes of the text content in the total number of requests
Requests per second: 133.36 [#/sec] (mean) #The number of concurrency supported per second, here is 133.36 concurrency supported per second
Time per request: 74.984 [ms] (mean) #The total time it takes to complete a request
Time per request: 7.498 [ms] (mean, across all concurrent requests) #The total time it takes to complete one request among all concurrent requests
Transfer rate : 12259.83 [Kbytes/sec] received #Total number of bytes received per second (KB)

#Test time result data, pay attention to the Total item, in the case of 10 concurrency, the minimum time to complete a request (min ) 9 milliseconds, maximum time consuming (max) 146 milliseconds
Connection Times (ms)
 min mean[+/-sd] median max
Connect: 0 0 0.7 0 20
Processing: 9 74 22.2 68 146
Waiting: 6 73 22.1 67 143
Total: 9 74 22.2 68 146

#Percentage of completed requests, for example, the first item, 50% 68, means that half of the requests are completed within 68 milliseconds, 95% 129, means that out of 1000 requests, 100% Ninety-eight requests were completed in 129 milliseconds
Percentage of the requests served within a certain time (ms)
 50% 68
 66% 78
 75% 86
 80% 92
 90% 109
 95% 122
 98% 129
 99% 133
 100% 146 (longest request)

ab option

Enter the following command to get the ab help document, which explains the usage and the meaning of command options respectively

D:phpStudyApachebin>ab -h

ab帮助文档的输出

Usage: ab [options] [http://]hostname[:port]/path
Options are:
    -n requests     Number of requests to perform
    -c concurrency  Number of multiple requests to make at a time
    -t timelimit    Seconds to max. to spend on benchmarking
                    This implies -n 50000
    -s timeout      Seconds to max. wait for each response
                    Default is 30 seconds
    -b windowsize   Size of TCP send/receive buffer, in bytes
    -B address      Address to bind to when making outgoing connections
    -p postfile     File containing data to POST. Remember also to set -T
    -u putfile      File containing data to PUT. Remember also to set -T
    -T content-type Content-type header to use for POST/PUT data, eg.
                    'application/x-www-form-urlencoded'
                    Default is 'text/plain'
    -v verbosity    How much troubleshooting info to print
    -w              Print out results in HTML tables
    -i              Use HEAD instead of GET
    -x attributes   String to insert as table attributes
    -y attributes   String to insert as tr attributes
    -z attributes   String to insert as td or th attributes
    -C attribute    Add cookie, eg. 'Apache=1234'. (repeatable)
    -H attribute    Add Arbitrary header line, eg. 'Accept-Encoding: gzip'
                    Inserted after all normal header lines. (repeatable)
    -A attribute    Add Basic WWW Authentication, the attributes
                    are a colon separated username and password.
    -P attribute    Add Basic Proxy Authentication, the attributes
                    are a colon separated username and password.
    -X proxy:port   Proxyserver and port number to use
    -V              Print version number and exit
    -k              Use HTTP KeepAlive feature
    -d              Do not show percentiles served table.
    -S              Do not show confidence estimators and warnings.
    -q              Do not show progress when doing more than 150 requests
    -l              Accept variable document length (use this for dynamic pages)

    -g filename     Output collected data to gnuplot format file.
    -e filename     Output CSV file with percentages served
    -r              Don't exit on socket receive errors.
    -m method       Method name
    -h              Display usage information (this message)

英文看不懂，附上中文解释，网上抄的，我没试过

-n 即requests，用于指定压力测试总共的执行次数。
-c 即concurrency，用于指定压力测试的并发数。
-t 即timelimit，等待响应的最大时间(单位：秒)。
-b 即windowsize，TCP发送/接收的缓冲大小(单位：字节)。
-p 即postfile，发送POST请求时需要上传的文件，此外还必须设置-T参数。
-u 即putfile，发送PUT请求时需要上传的文件，此外还必须设置-T参数。
-T 即content-type，用于设置Content-Type请求头信息，例如：application/x-www-form-urlencoded，默认值为text/plain。
-v 即verbosity，指定打印帮助信息的冗余级别。
-w 以HTML表格形式打印结果。
-i 使用HEAD请求代替GET请求。
-x 插入字符串作为table标签的属性。
-y 插入字符串作为tr标签的属性。
-z 插入字符串作为td标签的属性。
-C 添加cookie信息，例如："Apache=1234"(可以重复该参数选项以添加多个)。
-H 添加任意的请求头，例如："Accept-Encoding: gzip"，请求头将会添加在现有的多个请求头之后(可以重复该参数选项以添加多个)。
-A 添加一个基本的网络认证信息，用户名和密码之间用英文冒号隔开。
-P 添加一个基本的代理认证信息，用户名和密码之间用英文冒号隔开。
-X 指定使用的代理服务器和端口号，例如:"126.10.10.3:88"。
-V 打印版本号并退出。
-k 使用HTTP的KeepAlive特性。
-k 使用HTTP的KeepAlive特性。
-d 不显示百分比。
-S 不显示预估和警告信息。
-g 输出结果信息到gnuplot格式的文件中。
-e 输出结果信息到CSV格式的文件中。
-r 指定接收到错误信息时不退出程序。
-h 显示用法信息，其实就是ab -help。

Generally use -n, -c, -t
-n Total number of requests
-c Concurrency count
-t Concurrency count within specified seconds

Let me talk about the usage of

-t. The following means 10 concurrent visits in 20 seconds

D:phpStudyApachebin> ab -c 10 -t 20 http://example.com/phpinfo.php
This is ApacheBench, Version 2.3 <$Revision: 1604373 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache. org/

Benchmarking example.com (be patient)
Finished 3011 requests


Server Software: nginx/1.6.2
Server Hostname: example.com
Server Port: 80

Document Path: /phpinfo.php
Document Length: 94804 bytes

Concurrency Level: 10
Time taken for tests: 20.006 seconds
Complete requests: 3011
Failed requests: 495
 (Connect: 0, Receive: 0, Length: 495, Exceptions: 0)
Non-2xx responses: 54
Total transferred: 280896464 bytes
HTML transferred: 280344222 bytes
Requests per second: 150.50 [#/sec] (mean)
Time per request: 66.444 [ms] (mean)
Time per request: 6.644 [ms] (mean, across all concurrent requests)
Transfer rate: 13711.44 [Kbytes/sec] received

Connection Times (ms)
 min mean[+/-sd] median max
Connect: 0 0 0.5 0 10
Processing: 6 66 17.6 63 203
Waiting: 5 64 17.3 61 202
Total: 6 66 17.6 63 203 % 75
 90% 82
 95% 92
 98% 113
 99% 125
 100% 203 (longest request)


Copy code


Special note: If you visit http://www.example.com/, no specific File, remember to add the last /, be sure to add it! ! ! Otherwise, you may not be able to access it! ! !

Code complexity

Browser parsing

Web server configuration

Geolocation and network issues

If the web server is stored overseas, then we access the overseas website locally. As you can imagine, we pass through the router nodes, server nodes, and then through the optical cable under the Pacific Ocean, and then access the web server. After processing by the web server , continue with the submarine optical cable, and jump back to each server and routing node. This will definitely affect the test results (my suggestion is to conduct ab testing directly on the server)

Response file size

If a 3MB page is sent, the server network card will split the 3MB data into a single small data packet. During the transmission process, if only one data packet is damaged or lost, all data packets will be resent, so the sent packet The smaller the better, and at the same time, the smaller the data transfer, the faster it is transferred to the user's machine.

code complexity

The complexity of the code means complex processing of business logic, as well as file calls, database access, remote API interface calls, etc., which will affect the processing time

Browser parsing

Each browser has its own way of processing js, css, and html. Just think about the difference between IE8 and below and chrome

Web server settings

After the general Web server is installed, simple settings can be used to implement Web services (out of the box). However, such settings do not maximize the performance of the Web server. Senior engineers are required to configure the server to maximize performance. Here are simple and practical Keep-Alive instructions.

The function of Keep-Alive is that the web server opens a specific number of connections and keeps these connections open so that it can quickly process incoming requests, so that it does not open a connection for each incoming request and then process the request. , Reduce the number of processes on the server that process requests, thereby increasing the number of concurrencies.

Use -k in ab for testing, such as the following sentence

#-k, which allows the server to save 5 concurrent connections in an open and allowed state, allowing other requests to be processed quickly and reducing the processing time of creating new requests.
ab -c 5 -t 100 -k http://www.example.com/