The aim of this benchmark is to answer the question do you need framework for your tiny microservice.
Let's consider simple example where we need framework only to parse URL
(someone may decide that it's sufficient reason to start use framework for tiny microservice).
For gin example will use this implementation,
and for plain go example will use this implementation.
Code looks in next way:
Please run next command with purpose to make go web-server available
from one docker container to another docker container and especially for apachebench.
# OSX
sudo ifconfig lo0 alias 10.254.254.254
# Ubuntu
sudo ifconfig lo 10.254.254.254mkdir -p /tmp/benchmark/src/gin
curl -o /tmp/benchmark/src/gin/main.go \
"https://gist.githubusercontent.com/cn007b/318180624ee1bbe1c444439004cae761/raw/a9bceec28eb48b32551d595a1991b573b6ec6b48/gin.go"
docker run -it --rm -v /tmp/benchmark:/app -w /app -e GOPATH='/app' \
cn007b/go sh -c 'go get github.com/gin-gonic/gin && go install gin'mkdir -p /tmp/benchmark/src/plain
curl -o /tmp/benchmark/src/plain/main.go \
"https://gist.githubusercontent.com/cn007b/677973fc5f9d3fc75869fffff7c22bb8/raw/ebb63bd7d49ccab10a5b9a3735c7c4d80686763a/plain.go"
docker run -it --rm -v /tmp/benchmark:/app -w /app -e GOPATH='/app' \
cn007b/go sh -c 'go install plain'# run gin
docker run -it --rm -p 8080:8080 -v /tmp/benchmark:/app -w /app -e GOPATH='/app' \
cn007b/go sh -c './bin/gin'
# run apachebench
docker run -ti --rm cn007b/ubuntu ab -k -n 5000 -c 100 -t 2 "http://10.254.254.254:8080/v1/id/7"
# run plain go
docker run -it --rm -p 8080:8080 -v /tmp/benchmark:/app -w /app -e GOPATH='/app' \
cn007b/go sh -c './bin/plain'
# run apachebench
docker run -ti --rm cn007b/ubuntu ab -k -n 5000 -c 100 -t 2 "http://10.254.254.254:8080/v1/id/7"Finished 7637 requests
Server Software:
Server Hostname: 10.254.254.254
Server Port: 8080
Document Path: /v1/id/7
Document Length: 10 bytes
Concurrency Level: 100
Time taken for tests: 2.000 seconds
Complete requests: 7637 ‼️
Failed requests: 0
Keep-Alive requests: 7637
Total transferred: 1199009 bytes
HTML transferred: 76370 bytes
Requests per second: 3818.42 [#/sec] (mean) ‼️
Time per request: 26.189 [ms] (mean)
Time per request: 0.262 [ms] (mean, across all concurrent requests)
Transfer rate: 585.44 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 1 8.3 0 129
Processing: 1 25 15.1 24 155
Waiting: 1 25 15.1 24 155
Total: 1 26 17.3 24 196
Percentage of the requests served within a certain time (ms)
50% 24
66% 28
75% 31
80% 33
90% 41
95% 56
98% 73
99% 90
100% 196 (longest request)
Finished 15349 requests
Server Software:
Server Hostname: 10.254.254.254
Server Port: 8080
Document Path: /v1/id/7
Document Length: 11 bytes
Concurrency Level: 100
Time taken for tests: 2.000 seconds
Complete requests: 15349 ‼️
Failed requests: 0
Keep-Alive requests: 15349
Total transferred: 2333048 bytes
HTML transferred: 168839 bytes
Requests per second: 7672.79 [#/sec] (mean) ‼️
Time per request: 13.033 [ms] (mean)
Time per request: 0.130 [ms] (mean, across all concurrent requests)
Transfer rate: 1138.93 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 5.9 0 123
Processing: 1 13 2.9 12 49
Waiting: 1 13 2.9 12 49
Total: 1 13 6.7 12 142
Percentage of the requests served within a certain time (ms)
50% 12
66% 13
75% 14
80% 15
90% 16
95% 17
98% 20
99% 23
100% 142 (longest request)
100% sure you've answered the main question of this benchmark! 🙂
And now you have clue how much will you pay for syntax sugar which performs URL parsing.
You can find more stuff like this in my demo repo.
Source code, examples, explanation info, etc. just go and check it out!
