FEBCAM: A NOVEL EFFICIENT FILTER STRUCTURE BASED ON CAM AND BLOOM FILTER

Abstract

This paper proposes an efficient filtering structure, Feature extraction Bloom Content-Addressable Memory (FeBCAM), which integrates Content-Addressable Memory (CAM) with the Bloom filter (BF) to enhance membership query performance. The proposed FeBCAM structure supports adding, querying, and member-deletion operations; the last of these is novel compared with existing Bloom-based filters. Experimental evaluations demonstrate that FeBCAM achieves significant memory savings, i.e., 44% to 76%, and reduces filtering error rates by 81% to 87% compared with previous filter designs. However, these improvements come at the cost of higher query latency, with an average increase of approximately 1.8× to 2.6× compared with conventional Bloom-based filters. FeBCAM provides an efficient filtering solution for memory-constrained applications that require frequent filter-member updates.