Multi-station integrated power grid system has a large number of cloud and edge data centers. It needs to solve the problem of collaborative use of cloud, edge and terminal resources, realize the rapid migration of computing tasks in the case of failure, and achieve the consistency of primary and standby resources in the cloud and edge. It needs to solve the problem of streaming processing in high concurrent state. This paper studies the resource scheduling optimization technology adapted to the power cloud edge collaboration, innovatively proposes a multi-data center resource optimization and upgrading method based on the graph data structure, adapts to the multi-center resource optimization and upgrading scenario, uses the RDF resource description framework, TLGM data model to build the multi-data center resource database, uses the global scheduler, the edge scheduler to process the calculation request, uses the data linkage state data model, scheduling rules The probability calculation matrix converts the resource consistency and resource utilization into graph query, and uses the original graph retransmission, subgraph merging technology and efficient load balancing to realize graph query, so as to realize the optimization and upgrading of resources in multiple data centers. This paper innovatively proposes a stream data processing method that is suitable for the cloud edge collaborative multi- data center scenario. It is suitable for the cloud edge collaborative multi-data center stream data processing and analysis scenario. The stream business control and orchestration center construct a serial-parallel collaborative flow analysis process based on the pipeline processing model and parallel processing model. The flow control center control terminal, edge data center, and cloud data center implement flow analysis and scheduling according to the business priority, give full play to the advantages of cloud edge collaborative multi-data center distributed computing to achieve rapid processing and analysis of streaming data.
|