FastCGI-Bench - NodeJS vs .NET Core

The goal of this project is to demonstrate that a .NET Core web application can reach the performance of a NodeJS one if it uses only async techniques for handling input/output. Since I was unable to find a fully async FastCGI client library for .NET Core, I've developed one: AsyncFastCGI.NET.

These are the tested FastCGI libraries:

Project name	Github page	Framework	Language
AsyncFastCGI.NET	https://github.com/bolner/AsyncFastCGI.NET	.NET Core	C#
node-fastcgi	https://github.com/fbbdev/node-fastcgi	NodeJS	Javascript
LukasBoersma/FastCGI	https://github.com/LukasBoersma/FastCGI	.NET Core	C#

Benchmark results

The ApacheBench tool was used to measure the performance. As you can see, the non-async library fails during higher concurrency. All tests were run 10 times, and the best one (req/sec) was selected.

Concurrency: 20 simultanous connections / 20'000 requests

    ab -c 20 -n 20000 127.0.0.1/PATH

Library	Req. /sec	Req. Time	Conc. R.T.	Longest R.
AsyncFastCGI.NET	19494.64	1.026 ms	0.051 ms	18 ms
NodeJS	19453.72	1.028 ms	0.051 ms	13 ms
LB FastCGI	4249.07	4.707	0.235	3040 ms

Req. Time: mean | Conc. R.T.: mean, across all concurrent requests

Concurrency: 400 simultaneous connections / 200'000 requests

    ab -c 400 -n 200000 127.0.0.1/PATH

Library	Req. /sec	Req. Time	Conc. R.T.	Longest R.	Failed
AsyncFastCGI.NET	19893.88	20.107 ms	0.050 ms	2044 ms	0
NodeJS	21411.16	18.682 ms	0.047 ms	1062 ms	0
LB FastCGI	fails	fails	fails	fails	fails

Req. Time: mean | Conc. R.T.: mean, across all concurrent requests

Conclusions:

The .NET Core applications can compete with the NodeJS ones regarding performance of I/O operations.
Using FastCGI clients can be an alternative to the traditional ways of implementing web applications for .NET, which force too many constraints on the developers. The later being the biggest reason for their unpopularity.

Installation

A docker file is provided for setting up the test environment for comparing the performances of NodeJS and .NET Core FastCGI clients, using an Nginx webserver.

Create the docker image:
```
  docker/build_image.sh
```
Create and start a container. The parameter is the public http port.
```
  docker/create_container.sh 8080
```

  docker/login.sh
  or
  docker/root_login.sh

Nginx config

4 NodeJS and 2x4 C# processes accept FastCGI request on different ports.

    upstream fastcgi_backend_asyncfcgi {
            server 127.0.0.1:8080;
            server 127.0.0.1:8081;
            server 127.0.0.1:8082;
            server 127.0.0.1:8083;
    }

    upstream fastcgi_backend_nodejs {
            server 127.0.0.1:7070;
            server 127.0.0.1:7071;
            server 127.0.0.1:7072;
            server 127.0.0.1:7073;
    }

    upstream fastcgi_backend_lbfastcgi {
            server 127.0.0.1:9090;
            server 127.0.0.1:9091;
            server 127.0.0.1:9092;
            server 127.0.0.1:9093;
    }

    server {
            listen 80 default_server;
            listen [::]:80 default_server;

            root /var/www/html;
            server_name _;

            location / {
                    try_files $uri $uri/ =404;
            }

            fastcgi_keep_conn off;
            fastcgi_request_buffering off;

            location /asyncfastcgi {
                    include /etc/nginx/fastcgi_params;
                    fastcgi_pass fastcgi_backend_asyncfcgi;
            }

            location /nodejs {
                    include /etc/nginx/fastcgi_params;
                    fastcgi_pass fastcgi_backend_nodejs;
            }

            location /lbfastcgi {
                    include /etc/nginx/fastcgi_params;
                    fastcgi_pass fastcgi_backend_lbfastcgi;
            }
    }

bolner / FastCGI-bench