Design of Micro-service Architecture for Lightweight Dispatching and Control System

As the characteristics of integrated and coordinated operation of the current power grids become more significant, the SOA-based power grid dispatching control system has shown inadequacy in its ability to support information sharing and business collaboration between dispatch agencies. In recent years, micro-service architecture (MSA) has been widely deployed due to its low coupling of functional modules, strong scalability and performance advantages. At the same time, the development and progress of lightweight virtualization technologies such as containers (Docker) and hybrid storage architectures also provide low-level performance guarantees for MSA. Therefore, starting from the micro-service architecture, this paper designs and verifies a micro-service system architecture for lightweight dispatching and control system. The architecture is designed based on the MSA principle and realizes efficient resource scheduling and load balancing of the system through the container resource allocation method based on reinforcement learning. Meanwhile, the system supports real-time data access performance for the micro-service platform through complex heterogeneous data hybrid management technology and unified data access interface. The experimental results show that the system significantly improves the performance of the system under the sudden load, while effectively reducing the system delay and storage resource consumption.