Towards Efficient NDN Framework for Connected Vehicle Applications

Abstract

Named Data Networking (NDN) recently arises as a promising technology to support connected vehicle (CV) applications due to the match between their characteristics. However, the fast mobility and the vast number of vehicles raise great challenges in designing a scalable and efficient NDN network for CV applications. Therefore, in this paper, we develop an NDN based CV application framework that handles the challenge through innovations in two aspects. First, we propose a hierarchical hyperbolic NDN backbone architecture (H2NDN). H2NDN exploits the location dependency of CV applications to develop a hierarchical router topology and a hierarchical data/interest namespace. As a result, efficient and scalable data retrieval can be achieved by only configuring static forwarding information base (FIB) on NDN routers. To avoid overloading high-level routers, H2NDN integrates hyperbolic routing into the hierarchical architecture through carefully designed hyperbolic planes. Second, on top of the H2NDN architecture, we further model the optimal data caching problem. Based on the modeling, we propose a distributed adaptive caching strategy that can greatly improve the efficiency of the H2NDN backbone in supporting CV applications. Extensive ndnSIM based experiments with real traffic data in a city prove the efficiency and scalability of the proposed NDN application framework.