A Temporally Hierarchical Deployment Architecture for an Enhanced Name Resolution System

Abstract

The challenging requirements of the fifth generation (5G) cellular networks motivate the need to explore the feasibility of delivering services over new network architectures. Information-Centric Networking (ICN) is an emerging and promising network to satisfy 5G requirements. The name resolution is at the heart of ICN. We focus on satisfying the delay-sensitive requirement of the Name Resolution Service (NRS) in a 5G-ICN integrated network. We aim to design a local NRS that provides a deterministic low latency name resolution service. In this paper, we propose a temporally hierarchical deployment architecture for an Enhanced Name Resolution System (ENRS) to realize deterministic latency. The ENRS quantifiably organizes the nodes into hierarchical and nested domains by latency constraints. We design demand-aware name registration and resolution schemes to achieve constant forwarding hops in order to realize local resolution and forwarding locality. We introduce a tolerable latency-based peer resolver forwarding algorithm to improve the query hit ratio. We present a proactive name binding replicas distribution approach based on temporal–spatial features to reduce the resolution latency and query traffic. The video streaming monitoring service in Smart Home is used as a typical use case to show the continuity of service guaranteed by ENRS. Analysis demonstrates that ENRS can achieve deterministic latency. Evaluation results show that the average query hit ratio of ENRS outperforms the K-NearestNeighbor-Distributed Name Resolution System (KNN-DNRS) and Random Name Resolution System (Random-NRS) with 23.2% and 18.1%, respectively. The query traffic overhead of ENRS is up to 33.3 times smaller than KNN-DNRS. ENRS can process up to 21 GB/s name resolution traffic when the user nodes are in the magnitude order of 106.