Stochastic Pre-classification for SDN Data Plane Matching

摘要: The Software Defined Networking (SDN) approach has numerous advantages, including the ability to program network through simple abstractions, provide a centralized view of state, and respond changing conditions. One main challenges in designing SDN enabled switches is efficient packet classification data plane. As complexity applications increases, plane becomes more susceptible Denial Service (DoS) attacks, which can result increased delays loss. Accordingly, there strong need for architectures that operate efficiently presence malicious traffic. In particular, protect authorized flows from DoS attacks. this work we utilize probabilistic structure pre-classify traffic with aim decoupling likely legitimate by leveraging locality flows. We validate our examining fundamental application: software defined firewall. For application, architecture dramatically reduces impact unknown/malicious on established/legitimate explore effect stochastic pre-classification prioritizing classification. show how be used increase effective Quality (QoS) established reduce adversarial