Joint Optimization of Energy and Delay in NGSO Satellite Constellations: Cached Data Sharing Based on Spreading Dynamics

Zeyu Hu,Zhonghui Jiang,Jinming Zhao,Yong Li

DOI: https://doi.org/10.1109/tvt.2023.3241332

IF: 6.8

2023-01-01

IEEE Transactions on Vehicular Technology

Abstract:Edge caching has developed rapidly due to its ability to optimize both service delay and energy consumption significantly by reducing the traffic load in backhaul links. However, the mobility of edge devices will cause the reduction of the cached data available and the disappearance of the aforementioned advantages. This paper is intended for the cache strategy in the non-geostationary orbit (NGSO) satellite constellations and formulates an effective data sharing strategy between NGSO satellites for avoiding the invalidation of cached data. The density of cached data is a key indicator for the sharing strategy, because excessive sharing can increase the probability of cache hits at the expense of waste of resources and vice versa. Spreading dynamics is a physical model to describe the interaction of individuals in complex networks, which is helpful to predict the decisive threshold in the system. Therefore, this paper identifies the relationship between satellite speed, constellation structure, data size and hit probability, calculating a decisive threshold for the effective data density based on spreading dynamics. Meanwhile, the ideas of water-filling algorithm and reinforcement learning algorithm have been improved and applied to the data sharing strategy limited by the satellite cache space. Finally, the proposed strategy is compared with the traditional strategy in a variety of scenarios, and the results verify the effectiveness of the proposed optimization strategy in terms of energy consumption and service delay.

telecommunications,engineering, electrical & electronic,transportation science & technology