Interest Flow Control Method Based on User Reputation and Content Name Prefixes in Named Data Networking

摘要

Interest Flooding Attack (IFA) is a big problem in Named Data Networking (NDN). In IFA, an attacker repeats sending an excessive number of Interest packets requesting non-existing contents within short time in order to overload the network. It causes service disruptions for normal users. Pushback mechanism is a representative countermeasure against IFA in NDN. However, the mechanism also limits Interests from normal users, because it controls the flow in all routers affected by IFA. In addition, they assume only simple constant attack model in NDN. As a result, the data acquisition of normal users decreases. In this paper, we propose an Interest flow control method based on user reputation and content name prefixes in Named Data Networking, called ICRP. In ICRP, an edge router limits only Interests from malicious users who are attackers by user reputation. Here, reputation is the value that means the transmission degree of Interest requiring existing contents. As the reputation reflects the past behavior of each user, ICRP considers malicious users change their behavior. Furthermore, the edge router reduces the number of malicious Interests by content name prefixes. The edge router makes a blacklist of non-existing name prefixes requested by the detected malicious users. We evaluate ICRP by simulation. We confirm that ICRP can suppress the limitation to Interests from normal users. Furthermore, ICRP can alleviate the fluctuation the data acquisition rate of normal users even if malicious users change their behavior.