School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China

The construction of a novel power system has promoted massive diversification of the power information security service. Neither the single physical machine nor the cloud computing virtual machine can meet the security and timeliness requirements of the execution of each type of service tasks. Therefore, for the new hybrid underlying hardware pattern with the coexistence of the physical cryptographic machine and the virtual cryptographic machine in the information security service system of power Internet of Things, the system execution efficiency can be optimized under the constraint of ensuring the task security by coordinating the heterogeneous computing resources. First, a three-level double-mapping multi-task scheduling architecture for the information security service of power Internet of Things is designed, which can describe the mapping rules between multi-round interactive secure tasks and heterogeneous computing resources. Then, the execution label of the security service task is determined based on the weighted deep embedding clustering algorithm. Finally, the naive Bayes and quantum particle swarm optimization algorithms are used to complete the creation of computing nodes and the task mapping. The simulation results show that the proposed clustering mapping scheduling algorithm can use the task mapping rules to match the new-type heterogeneous security service computing resources. Compared with the two-stage scheduling algorithm and the non-dominated sorting genetic algorithm, the task safety factor of the proposed algorithm can be increased by up to 58.4%, and the total makespan can be shortened by up to 22.24%.