Django-doh
Django-doh is a simple Django app to serve a DOH (DNS Over HTTPS) Proxy. It resolves DNS query on HTTP.
Quick start
- Install via pip:
pip install django-doh
- Add
'doh_server'
to yourINSTALLED_APPS
setting like this:
INSTALLED_APPS = [
...
'doh_server',
]
- Include the
doh_server
URLconf in your project urls.py like this:
path('dns-query', include('doh_server.urls')),
- Add the conf in your settings like this:
DOH_SERVER = {
"RESOLVER": "internal",
# "RESOLVER": "8.8.8.8",
"AUTHORITY": "",
"LOGGER_LEVEL": "DEBUG",
}
To use the local resolver on the server where you run Django, use "internal"
.
Implementation
RFC 8484
Json implementation
Use with Firefox
in about:config
edit:
network.trr.mode;3
network.trr.uri;https://127.0.0.1/dns-query
For the URI, add your URI for your reverse proxy serving Django.
You can use Nginx as reverse proxy :
events {
}
http {
server {
listen 443 ssl http2 default_server;
location /dns-query {
proxy_set_header Host $host;
proxy_pass http://127.0.0.1:8000/dns-query;
}
ssl_certificate /etc/nginx/cert.pem;
ssl_certificate_key /etc/nginx/key.pem;
}
}
Test suite
Run using tox
.
tox --parallel
Benchmark
Macbook Pro 2019 Processor 2,4 GHz Intel Core i5 Memory 8 GB 2133 MHz LPDDR3
Server Django and a reverse proxy Nginx in a Docker container.
apib -c 100 -d 60 @benchmark_get_url.txt
HTTP/1.1
Duration: 60.032 seconds
Attempted requests: 7911
Successful requests: 7911
Non-200 results: 0
Connections opened: 104
Socket errors: 0
Throughput: 131.780 requests/second
Average latency: 734.647 milliseconds
Minimum latency: 27.760 milliseconds
Maximum latency: 9819.022 milliseconds
Latency std. dev: 710.929 milliseconds
50% latency: 606.548 milliseconds
90% latency: 1360.734 milliseconds
98% latency: 2845.927 milliseconds
99% latency: 3612.085 milliseconds
Client CPU average: 0%
Client CPU max: 0%
Client memory usage: 0%
Total bytes sent: 1.15 megabytes
Total bytes received: 2.44 megabytes
Send bandwidth: 0.15 megabits / second
Receive bandwidth: 0.33 megabits / second