Vehicle-To-Vehicle Distributed Storage In Vehicular Networks

Binbin Hu,Luoyang Fang,Xiang Cheng,Liuqing Yang

DOI: https://doi.org/10.1109/ICC.2018.8422220

2018-01-01

Abstract:In recent years, wireless communication demands from high-speed vehicles rapidly increase, due to the enrichment of various emerging vehicular applications including road safety, intelligent transportation, in-vehicle entertainments, self-driving, etc. Such high communication demands result in significant burdens on the supporting infrastructure in vehicular networks, which in turn deteriorate the quality of experience (QoE) of vehicular users. In the literature, local popular content caching is an effective approach to alleviate the burden on core networks as well as reducing the content downloading delay. However, the burden is simply shifted to RSUs if popular contents are cached at RSUs. In addition, the frequency reuse in space is also inadequate. Caching popular contents at vehicles may be an effective approach to solve the aforementioned problems. However, the critical issue of vehicle-based storage schemes is the high mobility nature of vehicles, which could lead to loss of stored data due to the storage vehicles' leaving the region of interests. To cope with the vehicle mobility issue, we propose a vehicle-based distributed storage scheme via local V2V communications in this paper. The key of the proposed data storage scheme is to maintain the survival of the stored data in a designated region by transferring the stored data from leaving vehicles to incoming vehicles at the entrance/exit of the region, in order to prevent data loss due to the high mobility of vehicles. Structured redundancy via erasure coding is also introduced in order to combat the volatile V2V links. Theoretical analysis and numerical results validate the effectiveness of our proposed vehicle-based distributed storage scheme and suggest that the stored data could survive for up to tens of hours under a typical road condition.