View of Research on a Deep Reinforcement Learning-Based Adaptive Replica Allocation Method for Data Replica Placement Optimization in Distributed Storage Systems | Journal of Computer Technology and Software