Needle in a Haystack: Mitigating Content Poisoning in Named-Data Networking

Abstract

Named-Data Networking (NDN) is a candidate next-generation Internet architecture designed to address some limitations of the current IP-based Internet. NDN uses the pull model for content distribution, whereby content is first explicitly requested before being delivered. Efficiency is obtained via routerbased aggregation of closely spaced requests for popular content and content caching in routers. Although it reduces latency and increases bandwidth utilization, router caching makes the network susceptible to new cache-centric attacks, such as content poisoning. In this paper, we propose a ranking algorithm for cached content that allows routers to distinguish good and (likely) bad content. This ranking is based on statistics collected from consumers actions following delivery of content objects. Experimental results support our assertion that the proposed ranking algorithm can effectively mitigate content poisoning attacks.