Naxos: A Named Data Networking Consensus Protocol

As a promising network model different from TCP/IP, Named Data Networking (NDN)still lacks consensus protocol support which is a fundamental building block to establish fault-tolerant systems and other distributed applications requiring strong consistency. However, implementing consensus protocols in NDN faces severe important challenges. In this paper, we propose a Paxos-based consensus protocol, namely Naxos, to guarantee both performance by exploiting NDN features. Naxos adopts a Self-learning mechanism to improve the performance. We evaluate this new protocol by designing and implementing a replicated state machine on top of it. The results show that Naxos can achieve 50%∼69% commit latency improvement in wide area and 4.7∼5.57 times throughput improvement in a cluster over an intuitive Paxos implementation in NDN.