Elastic-C: Elastic Sketch Using Counter-Based Swapping

초록

In order to keep the network environment smooth, it is important to identify the causes of problems such as congestion control, DoS attacks, and scan attacks in the data stream. To solve this problem, it is required to measure the frequency of each flow of the data stream to find the flow causing the problem, but it is difficult to accurately and quickly measure the data stream in a limited environment. Unlike existing algorithms, the recently announced Elastic Sketch uses a hash table in front of a counter and uses the expulsion method by frequency voting (hereinafter, voting) to separately store high-frequency flows in the hash table. As a result, since the measurement error that can be caused by a counter with a high frequency flow is reduced, a significant number of flows with a small frequency become accurate. However, since the voting method they used is an algorithm that is performed on a hash table, several problems arise, such as requiring additional memory and operation. In this paper, we analyze the problems caused by the voting algorithm and introduce Elastic-C, a new method that can solve them. © 2021, Korean Institute of Communications and Information Sciences. All rights reserved.

키워드