Smart Name Lookup for NDN Forwarding Plane via Neural Networks

Abstract

Name lookup is a key technology for the forwarding plane of content router in Named Data Networking (NDN). To realize the efficient name lookup, what counts is deploying a high-performance index in content routers. So far, the proposed indexes have shown encouraging performance, most of which are optimized for or evaluated with URLs collected from the Internet, as the large-scale NDN names are not available yet. However, the performance of indexes heavily relies on the distributions of data indexed. The distributions of URLs indexed in memory may differ from those of names independently generated by contentcentric applications online. Therefore, these indexes will have to be redesigned to adapt to the distributions of the future real NDN names so as to ensure the performance. Focusing on this gap, a smart mapping model via neural networks is proposed to build an index LNI for NDN forwarding plane. Through learning the distributions of the names indexed in the static memory, LNI can not only reduce the memory consumption and the probability of false positive, but also ensure the performance without redesign in future real name lookup. Experimental results show that LNIbased FIB can reduce the memory consumption to 58.258 MB for 2 million names. Moreover, as it can be deployed on SRAMs, the throughput is about 177 MSPS, which well meets the current network requirement for fast packet processing.