DISCO: Memory Efficient and Accurate Flow Statistics for Network Measurement

Abstract

A basic task in network passive measurement is collecting flow statistics information for network state characterization. With the continuous increase of Internet link speed and the number of flows, flow statistics has become a great challenge due to the demanding requirements on both memory size and memory bandwidth in measurement devices. In this paper, we propose a DIScount COunting (DISCO) method, which is designed for both flow size and flow volume counting. For each incoming packet of length l, DISCO increases the corresponding counter assigned to the flow with an increment that is less than l. With an elaborate design on the counter update rule and the inverse estimation, DISCO saves memory consumption while providing an accurate unbiased estimator. The method is evaluated thoroughly under theoretical analysis and simulations with synthetic and real traces. The results demonstrate that DISCO is more accurate than related work given the same counter size. DISCO is also implemented on network processor Intel IXP2850 for performance test. Using only one MicroEngine (ME) in IXP2850, the throughput can reach up to 11.1Gbps under a traditional traffic pattern, and it increases almost linearly with the number of MEs employed.