Traffic archival, an indispensable task for network analysis, has suffered from a huge amount of data. These rapid growing data can exceed storage capacity and thwart real-time analysis. In order to take insight from Internet traffic, bitmap indexing is applied in this field. However, raw bitmap indexes can consume more space and longer delay for loading larger indexes into memory. This leads to the invention of various bitmap index compression algorithms to save storage and ensure query efficiency. This paper proposes a new algorithm called common affix merging with partition. By merging common affixes in a bit sequence, it saves more storage and conducts at least one order of magnitude faster bitwise operations when compared with ROARING, WAH, CONCISE, and COMPAX. In practice, besides bitwise operations, a query should also contain the operations of loading index files into memory. The experimental results demonstrate that the speedup is still considerable even taking loading time into consideration.
