A prediction-based detection algorithm against distributed denial-of-service attacks

摘要: Denial-of-Service (DoS) attacks especially distributed DoS (DDoS) have become significant and increasing threats to the Internet. Huge efforts from both academia industry been made on detection defense of DDoS attacks. However, most schemes do not directly aim at protecting victim itself (e.g., servers) but attack sources or intermediate network units. Although locating identifying attacking are critical stop for legal procedure, rapid efficient predicting happen in server is more important reduce damage caused by even prevent happening. this part has addressed sufficiently literature. In paper, we first briefly review research attacks, then discuss a method define quantify severs based available service rates. This because often direct one-point failure entire system. No matter whether there undergoing, if sever overloaded normal requests, effect imposed system equivalent that A prediction rate protected proposed, which applies Auto Regressive Integrated (ARIMA) model. Finally, investigate proposed predict through simulation studies with NS2. The results show algorithm effective