In the paper, the DDoS scenario is modelled as a mathematical programming problem. The defender strategically utilizes the limited resources to maximize the legitimate traffic, and he can adopt packet marking to observe the network status. The information extracts from the marking field can help the defender develop a defense strategy which combines packet filtering and packet blocking. A Lagrangean relaxation-based algorithm is proposed to optimally solve the problem.
Lin, Frank Yeong-Sung; Chen, Pei-Yu; and FanChiang, Chun-Wei, "Near-Optimal Defense Strategies against DDoS Attacks Based upon Packet Filtering and Blocking Enabled by Packet Marking" (2010). PACIS 2010 Proceedings. 167.