Skip to content

anuragkhanra/extra-tunnel

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Tunnel web server from private IP.

A tunneling system, where the tunnel acts both as a middle-man and an HTTP server. This enables users to access an HTTP server running locally, through a public-ip tunnel server, which can be hosted on a cloud server, like Heroku. The tunnel also supports channels, other than HTTP which enables users to access TCP servers, like SSH/FTP, running locally.

The system has 3 parts:

  • Tunnel: acts as the tunnel server
  • Server: enables local server to be hosted through Tunnel
  • Client: enables local clients to request through Tunnel

Think of Tunnel like a school. It has multiple channels, like a school has multiple classrooms. Each channel has a Server, like each classroom has a class teacher.. Any number of Clients can connect to a channel and send requests to the Server, and so can any number of students in a classroom ask questions to their class teacher.

Setup

Tunnel

In order to start, we need a Tunnel first. Let's set it up:

  1. Get Tunnel to your GitHub.
    1. Create an account on GitHub.
    2. Goto extra-tunnel repository, and fork it.
  2. Create Tunnel application in cloud.
    1. Create an account on Heroku.
    2. On Heroku dashboard, create a new app, like tunnelwebapp.
    3. Select the Deploy tab, and the choose GitHub as deployment method.
    4. In Connect to GitHub, type in extra-tunnel and Connect.
    5. In Manual Deploy, Deploy Branch when master is selected.

Server/Client

We need to install extra-tunnel locally in order to use it as Server or Client (a private-ip Tunnel would only be useful in testing).

# to use from command line
npm install -g extra-tunnel

# to use from node.js
npm install extra-tunnel



Usage

Host local HTTP server

Assuming your Heroku app name is tunnelwebapp, and your local HTTP server is running on port 80. The following command starts up a Server, which acts as a bridge between your local server localhost:80 and the Tunnel tunnelwebapp. Try opening https://tunnelwebapp.herokuapp.com in your browser, after running this command.

etunnel server --tunnel tunnelwebapp.herokuapp.com --server 80

Host local SSH server

All channels other than default / for HTTP are disabled by default. Lets enable it first by going to Tunnel setting on Heroku:

  1. Goto Heroku dashboard, and then choose Settings tab.
  2. In Config Variables, we need to add one, so select Reveal Config Vars.
  3. Set Key as KEYS_SSH, and Value as admin (or whatever you want).
  4. Select Add, this restarts the app with new config.
  5. You can see app logs at More -> View Logs.

Now that we have setup the key for /ssh channel, it is enabled and we are ready to setup the server. Assuming your Heroku app name is tunnelwebapp, and your local SSH server is running on port 22. The following command starts up a Server, which acts as a bridge between your local server localhost:22 and the Tunnel tunnelwebapp, on channel /ssh.

etunnel server -t tunnelwebapp.herokuapp.com -s 22 --channel /ssh --key admin

The common use of SSH is to access the terminal of a remote computer. In our case, since we are using Tunnel, we would now be able to access it, not just from LAN, but from anywhere in the world (with an internet connection). Unlike HTTP however, Tunnel is unable to act as an SSH server and hence you cannot connect directly to it with your SSH client.

To solve this problem, we have a Client. Any number of Clients can connect to a channel on the Tunnel. So, on a separate machine, install extra-tunnel using the command npm install -g extra-tunnel, and then start Client using the following command:

etunnel client -t tunnelwebapp.herokuapp.com -c 22 -n /ssh



Concept

Tunnel

It acts as a server on a single port, and manages communication between Clients and Servers through channels. Each Server registers to a unique channel (like / or /ssh), and any number of Clients can then connect to the Tunnel on that channel. The Tunnel also itself acts as a client on channel / forwarding any HTTP requests it receives on its port to the Server registered to channel /.

Server

It connects to the Tunnel, and registers to a unique channel using a key and a token. The key must match the one stored on the Tunnel for that channel. Once registered, the token is used to accept Clients. Server then acts a multiple local clients for forwarding requests to local server from specified channel, thus making you feel as if the Clients are running locally (even if its not). A Server registered to channel / will also receive HTTP requests from Tunnel, becuase Tunnel also acts as a Client to channel /.

Client

It connects to the Tunnel, and subscribes to a channel using a token. This token must match the one provided by the Server registered to this channel. Client then acts as a local server for forwarding requests of local clients to specified channel, thus making you feel as if the Server is running locally (even if its not). Any Client can also register to channel /, but this is unnecessary since you can directly request the Tunnel server instead.

Reference

Command Line

$ etunnel [<mode>] [options]

# mode: this is 'tunnel', 'server', or 'client'
# -t | --tunnel: address of tunnel
# -s | --server: address of server
# -c | --client: address of client
# -n | --channel: channel to register/subscribe
# -k | --key: key for registering server
# -o | --token: token for subscribing client
# -i | --ping: ping period to Tunnel
# -e | --keys: JSON object with keys of channels
# --keys_ch1: key for channel /ch1
# --keys_ch1_ch2: key for channel /ch1/ch2
# --version: get version
# --help: get this help
# environment variables are also accepted
# PORT: port number for tunnel
# TUNNEL: address of tunnel
# SERVER: address of server
# CLIENT: address of client
# CHANNEL: channel to register/subscribe
# KEY: key for registering server
# TOKEN: token for subscribing client
# PING: ping period to Tunnel in ms
# KEYS: JSON object with keys of channels
# KEYS_CH1: key for channel /ch1
# KEYS_CH1_CH2: key for channel /ch1/ch2

Node.js

const Tunnel = require('extra-tunnel');

Tunnel([<prefix>], [<options>]);
Tunnel.<Server/Client>([<prefix>], [<options>]);
// prefix: name of object in logs
// the following options are accepted (object)
// tunnel: address of tunnel ('localhost:7000')
// server: address of server ('localhost:7001')
// client: address of client ('localhost:7002')
// channel: channel to register/subscribe ('/')
// key: key for registering server ('')
// token: token for subscribing client ('')
// ping: ping period to Tunnel in ms (8000)
// keys: keys for each allowed channel ({'/': ''})



Cartoon Merferry

Listen: Epic Mountain; Inspiration: Samy Kamkar; Docs: Net Node.js; Play: Scrap Mechanic;
ASCII art: Taag; Reference: Markdown Cheatsheet; Prof. icons: Icon Experience

Releases

No releases published

Packages

No packages published

Languages

  • JavaScript 98.5%
  • Shell 1.4%
  • Batchfile 0.1%