Accurate Heavy Hitter Detection with Sliding Window Approach

Abstract

Network measurement is essential to many network applications. Heavy Hitter detection is a significant function within the network measurement algorithms. However, the interval method in traditional Heavy Hitter detection is not as accurate as of the sliding window method which can detect the Heavy Hitter that appears on the boundaries of different intervals. Accordingly, this paper proposes accurate Heavy Hitter detection, which combines the sliding window approach with sketch algorithms. Compared with the interval method, our experiments show that the sliding window improves the accuracy by 121.8x, 2.8x, and 3.5x on average for CM, CU, and Hashpipe, respectively.