A novel hybrid distributed storage strategy for Space Information Network
Bo Kong, Zhidong Xie, Dongming Bian, Gengxin Zhang, Jing Fang Hu · 2016
Hierarchical autonomous system (AS) network model makes the control of the entire Space Information Network (SIN) easier. In this paper, we investigate the distributed data storage problem in the SIN, a novel hybrid distributed storage (HDS) strategy is proposed to improve the data reliability based on the AS network model. After a data object is optimal encoded using maximum distance separable (MDS) codes, the storage process is divided into two phases: the inter-AS storage phase and intra-AS storage phase, respectively. The inter-AS storage is modeled as an optimal storage allocation problem for a heterogeneous network under multiple constraints, where different AS networks have different capacities and availabilities. Then, we deduce the distributed storage within each AS network to a data persistence problem. As a result, the data object is disseminated over the SIN and it can be recovered by collecting a few encoded packets. The proposed strategy utilizes the advantage of AS network model of SIN, while aiming to improve data reliability and energy efficiency.