Poster: Compressing Flow Statistic Counters

With the continuous increase of Internet link speed and the number of flows, flow statistics has become a challenging task due to the demanding requirements on both memory size and memory bandwidth. In this paper, we propose a DIScount COunting (DISCO) method, which is designed for both flow size and flow bytes counting. For each incoming packet, DISCO increases the corresponding counter assigned to the flow with an increment that is less than the actual packet length. With elaborate design of the counter update rule and the inverse estimation, DISCO saves large memory consumption while providing an accurate unbiased estimator. Experiments under real trace and synthetic traces demonstrate that DISCO is more accurate than related work given the same counter sizes for flow statistics.