AI Driven Security
Before now, the approach to handling DDoS attacks has always been to detect requests too frequent to be human, and block the ip addresses behind them. This approach is sorely limited for many reasons, one of which is that with the introduction of PWA technologies like React and Ember.js, it is actually completely possible for a non-malicious user to flood your server with requests in too short a period of time.
With Casper however, once a user hits your site, Casper starts tracking his requests and starts comparing them to its repertoire of data on what is ‘malicious’. For instance, it is suspicious if a user is requesting the same data 200 times in one second. On detecting this, the user is not blocked, but Casper changes their priority level, their requests still go through but much much slower. Meanwhile his activities are compared more aggressively against existing data on unusual browsing behavior. At this stage the user continues to browse the website seamlessly - just slower. If this artificial intelligence algorithms rate the user after further evaluation as ‘malicious’, his requests are then deprioritized even further to the point where his or her requests take minutes to be served.
This means that while the attacker continues to think he is hacking you, he is actually just killing his time while your real users continue to enjoy your services.