State Grid Corporation of China is now carrying out the construction of the full-service unified data center of electric power system to achieve the unified storage and management analysis of the data in the full-service scope

all data types and all time dimensions. While pushing the full-service business data

especially during the influx of burst mass data

the processing capacity of the full-service unified data center will face a great challenge. The scheduling method of data processing task is vital to improve the data processing capacity of the data center. Based on the analysis of the design scheme and architecture for the full-service unified data center of electric power system

a control scheduling model for the data processing task is given

and the priority of the task is analyzed synthetically from both the time and effective value of task. A multi-level queue based dynamic real-time priority scheduling(DRPS)algorithm is proposed

and the dynamically changing trend of priority for different types of tasks is analyzed. The experiment of data extract-transform-load(ETL)process of full-service unified data center is simulated on the CloudSim platform. The result shows that compared with the dynamic value-density(DVD)algorithm and the earliest deadline first(EDF)algorithm

the DRPS algorithm has a significant reduction in the deadline miss rate

and the completion rate of the task is significantly improved.