Improving denial of service resistance using dynamic local adaptations

We improve the resistance of gossip-based multicast to (Distributed) Denial of Service (DoS) attacks using dynamic local adaptations at each node. Each node estimates the current state of the attack on the system, and then adapts its behavior according to this local estimation. The adaptation is achieved through modeling the problem of propagating messages under a DoS attack as an optimization problem, and solving it using linear programming, independently at each node. Simulation results show that when the system is under attack, the local decisions each node takes bring the system to a stable point, which is the solution of the linear programming problem. The adaptation leads to propagation times that are 30% faster than those of existing DoS-resistant gossip-based protocols. Categories and Subject Descriptors C.4 [Performance of Systems]: Reliability, availability, and serviceability; C.2.4 [Computer-Communication Networks]: Distributed Systems--Distributed applications Keywords Ad...

• Type: Journal
• Year: 2008
• Where: SAC
• Authors: Gal Badishi, Idit Keidar
• Keywords: Applied Computing, Dynamic Local Adaptations, Linear Programming, Node, SAC 2008