Popularity-aware back-tracing partition cooperative cache distribution for space-terrestrial integrated networks

Space-terrestrial integrated networks consisting of low earth orbit (LEO) satellites and terrestrial users are widely developed for potentially diversified requirements of content distribution. With an obviously time-varying topology, however, designing a distribution strategy faces several explicit challenges, such as prolonged content access latency and significant transmission overheads, due to lack of contact opportunities and limited on-board storage space. In this study, therefore, a novel back-tracing partition directed on-path caching distribution mechanism (BPDM) is proposed for the file distribution in the hybrid LEO constellation and terrestrial network. In the proposed strategy, a group of feasible on-path cache nodes is iteratively selected by utilising a well-designed cross-timeslot graph, as well as a collaborative cached content placement strategy, called as multiple regions cooperative cache algorithm, by carefully considering diversified popularity of target files. As a result, the proposed BPDM could efficiently reduce redundant transmissions of content access for different users by fetching objective file mainly from limited quantities of intermediate caching nodes. Through the simulation results, the proposed method can obviously decrease the holistic overheads and access delay compared with the minimum spanning tree algorithm and Network Central Location (NCL) node selection metric.