Reducing the number of Bloom filters

Bloom Filters have been applied in many fields, including data base, network management, computer network, and computer communication etc., owing to its fast membership query and memory efficiency. In network field, Bloom Filters were used to lookup the forwarding tables in backbone routers and Data Centre switches. These solutions all build many Bloom Filters to achieve fast lookup. The main shortcoming of these solution is: too many Bloom Filters require too many hash computations and memory accesses, and the variety of Bloom Filter sizes poses challenges to system design and probably degrades the system performance. To address this issues, we proposed Set absorption algorithm to reduce the number of Bloom Filters, while balancing the size of Bloom Filters. Experimental results showed that after using our algorithm, better performance of the Bloom Filter-based solutions was obtained.