I have been learning node.js and socket.io lately. My question is how can I protect server against client side attacks?
This is my server code
io.sockets.on('connection', function (socket) { //users.push(socket); socket.on('message', function (data) { socket.on('disconnect', function () { }); socket.on('bcast', function (data) { socket.emit('news', { 'data': data }); socket.broadcast.emit('news', { 'data': data }); }); socket.on('login', function(data){ socket.emit('login', {'data': [ socket.id, data ] }); }); }); });
For example if a client uses chrome developer tools to do the flowing code
for(var i = 0; i<99999999999; i++) { socket.emit('bcast', {data: 'XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX'}); }
its going to kill the server.
Avoid DOS attacks by explicitly setting when a process should crash. Otherwise: This is just an educated guess: given many Node. js applications, if we try passing an empty JSON body to all POST requests — a handful of applications will crash.
There are several approaches you can take to defend against a DDoS attack: Black-holing or sinkholing: This approach blocks all traffic and diverts it to a black hole, where it is discarded. The downside is that all traffic is discarded -- both good and bad -- and the targeted business is taken off-line.
Defending Against DDoS Attacks It is much harder to respond after an attack is already under way. While DDoS attacks can't be prevented, steps can be taken to make it harder for an attacker to render a network unresponsive.
Botnets, which are vast networks of computers, can be used to wage DDoS attacks. They are usually composed of compromised computers (e.g., IoT devices, servers, workstations, routers, etc.), or zombies, that are controlled by a central server. Attackers don't necessarily need a botnet to conduct a DDoS attack.
Look into JS event throttling and debouncing!
Those techniques will help you prevent and detect attacks to a certain point (which is, in my opinion, enough for a small multiplayer socket game)...
EDIT:
In this jsfiddle: http://jsfiddle.net/y4tq9/9/
var sIO = {}; sIO.on = (function(){ var messages = {}; var speedLimit = 5; //5ms return function(message, handler) { messages[message] = messages[message] || {}; if(messages[message].timestamp && new Date().getTime() - messages[message].timestamp < speedLimit) return false; else messages[message].timestamp = new Date().getTime(); handler(); return true; //execute code, Ex: } }());
you can see that every request sent faster than 5ms will return false, otherwise the handler get run.
You simple disconnect the sockets who send request faster than 5ms (or 2ms, or 3ms depending on your network and your application's weight...).
You might as well using js event throttling on client site to make sure all of your requests doesn't send faster than the speed limit!
This technique will not provide absolute protection from exploiting, but it will prevent your server from crashing when attackers try to Dos...
The rate-limiter-flexible Node.js package can be used against DDoS attacks.
const { RateLimiterMemory } = require('rate-limiter-flexible'); const rateLimiter = new RateLimiterMemory({ points: 5, // 5 points duration: 1 // per second }); socket.on('bcast', data => { rateLimiter.consume(uniqueSocketId) // consume 1 point per event .then(() => { socket.emit('news', { 'data': data }); socket.broadcast.emit('news', { 'data': data }); }) .catch(rejRes => { // no available points to consume // emit error or another workaround }); });
Any event will be blocked, if it happens more than 5 times per second.
There is also the option for distributed applications using Redis. And some flexible settings like insurance and block strategy makes rate-limiter-flexible
highly available and fast.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With