Comparison of Cuckoo Hash Table and Bloom Filter for Fast Packet Filtering Using Data Plane Development Kit

Abstract

This paper presents a performance comparison of the cuckoo hash table and Bloom filter for packet classification with application in packet filtering with emphasis on lawful interception systems. Alongside enabling fast packet capture on commodity hardware, Data Plane Development Kit offers those methods already implemented in its library. We have chosen the cuckoo hash table as an exact classification method and the Bloom filter as a space-efficient probabilistic method that should have a better cache utilization and therefore more viable usage in performance critical systems, but allows false positive results. The execution time and memory requirements of the two methods are compared to determine whether the performance benefits of using a probabilistic method outweigh its drawbacks, and in which context those benefits arise.