python常用web框架性能测试(django,flask,bottle,tornado)

测了一下django、flask、bottle、tornado 框架本身最简单的性能。对django的性能完全无语了。

django、flask、bottle 均使用gunicorn+gevent启动，单进程，并且关闭DEBUG，请求均只返回一个字符串ok。

tornado直接自己启动，其他内容一致。

测试软件为 siege，测试os为cenos6 64位，测试命令为:

siege -c 100 -r 100 -b http://127.0.0.1:5000/

1	siege -c 100 -r 100 -b http://127.0.0.1:5000/

django测试结果为:

Transactions: 10000 hits
Availability: 100.00 %
Elapsed time: 18.51 secs
Data transferred: 0.02 MB
Response time: 0.18 secs
Transaction rate: 540.25 trans/sec
Throughput: 0.00 MB/sec
Concurrency: 99.35
Successful transactions: 10000
Failed transactions: 0
Longest transaction: 0.30
Shortest transaction: 0.12

1 2 3 4 5 6 7 8 9 10 11 12

Transactions: 10000 hits Availability: 100.00 % Elapsed time: 18.51 secs Data transferred: 0.02 MB Response time: 0.18 secs Transaction rate: 540.25 trans/sec Throughput: 0.00 MB/sec Concurrency: 99.35 Successful transactions: 10000 Failed transactions: 0 Longest transaction: 0.30 Shortest transaction: 0.12

django(去掉所有middleware)测试结果为:

Transactions: 10000 hits
Availability: 100.00 %
Elapsed time: 12.97 secs
Data transferred: 0.02 MB
Response time: 0.13 secs
Transaction rate: 771.01 trans/sec
Throughput: 0.00 MB/sec
Concurrency: 99.41
Successful transactions: 10000
Failed transactions: 0
Longest transaction: 0.28
Shortest transaction: 0.12

1 2 3 4 5 6 7 8 9 10 11 12

Transactions: 10000 hits Availability: 100.00 % Elapsed time: 12.97 secs Data transferred: 0.02 MB Response time: 0.13 secs Transaction rate: 771.01 trans/sec Throughput: 0.00 MB/sec Concurrency: 99.41 Successful transactions: 10000 Failed transactions: 0 Longest transaction: 0.28 Shortest transaction: 0.12

flask测试结果为:

Transactions: 10000 hits
Availability: 100.00 %
Elapsed time: 5.47 secs
Data transferred: 0.02 MB
Response time: 0.05 secs
Transaction rate: 1828.15 trans/sec
Throughput: 0.00 MB/sec
Concurrency: 96.25
Successful transactions: 10000
Failed transactions: 0
Longest transaction: 0.11
Shortest transaction: 0.00

1 2 3 4 5 6 7 8 9 10 11 12

Transactions: 10000 hits Availability: 100.00 % Elapsed time: 5.47 secs Data transferred: 0.02 MB Response time: 0.05 secs Transaction rate: 1828.15 trans/sec Throughput: 0.00 MB/sec Concurrency: 96.25 Successful transactions: 10000 Failed transactions: 0 Longest transaction: 0.11 Shortest transaction: 0.00

bottle测试结果为:

Transactions: 10000 hits
Availability: 100.00 %
Elapsed time: 4.55 secs
Data transferred: 0.02 MB
Response time: 0.04 secs
Transaction rate: 2197.80 trans/sec
Throughput: 0.00 MB/sec
Concurrency: 96.81
Successful transactions: 10000
Failed transactions: 0
Longest transaction: 0.09
Shortest transaction: 0.00

1 2 3 4 5 6 7 8 9 10 11 12

Transactions: 10000 hits Availability: 100.00 % Elapsed time: 4.55 secs Data transferred: 0.02 MB Response time: 0.04 secs Transaction rate: 2197.80 trans/sec Throughput: 0.00 MB/sec Concurrency: 96.81 Successful transactions: 10000 Failed transactions: 0 Longest transaction: 0.09 Shortest transaction: 0.00

tornado测试结果为:

Transactions: 10000 hits
Availability: 100.00 %
Elapsed time: 7.06 secs
Data transferred: 0.02 MB
Response time: 0.07 secs
Transaction rate: 1416.43 trans/sec
Throughput: 0.00 MB/sec
Concurrency: 99.51
Successful transactions: 10000
Failed transactions: 0
Longest transaction: 0.09
Shortest transaction: 0.01

1 2 3 4 5 6 7 8 9 10 11 12

Transactions: 10000 hits Availability: 100.00 % Elapsed time: 7.06 secs Data transferred: 0.02 MB Response time: 0.07 secs Transaction rate: 1416.43 trans/sec Throughput: 0.00 MB/sec Concurrency: 99.51 Successful transactions: 10000 Failed transactions: 0 Longest transaction: 0.09 Shortest transaction: 0.01

可见纯框架自身的性能为:

bottle > flask > tornado > django

结合实际使用:

tornado 使用了异步驱动，所以在写业务代码时如果稍有同步耗时性能就会急剧下降；
bottle需要自己实现的东西太多，加上之后不知道性能会怎样；
flask性能稍微差点，但周边的支持已经很丰富了；
django就不说了，性能已经没法看了，唯一的好处就是开发的架子都已经搭好，开发速度快很多

因为最近正在为一个项目选型发愁，所以就测了一下，记录在此吧。

PS: 2014-6-23 使用 centos6 64位 重新进行了测试，得出与生产环境更匹配的结果，并修改了文章。