DER: energy-aware financial task scheduling in multi-cloud environments with privacy constraints.

Area of Science:

• Computer Science
• Cloud Computing
• Artificial Intelligence

Background:

• Multi-cloud financial workflow execution faces challenges in balancing makespan, energy consumption, communication overhead, and privacy.
• Existing schedulers often fail to consider heterogeneous DVFS states of resources and varying task privacy levels.

Purpose of the Study:

• To develop a differential evolution-based scheduler (DER) for multi-cloud financial workflows.
• To address the four-objective optimization problem considering execution time, energy, communication cost, and privacy satisfaction.

Main Methods:

• Formulated a four-objective optimization model for scheduling financial workflows.
• Developed DER scheduler incorporating DVFS-based energy minimization, privacy-feasible resource assignment, and a hybrid mutation strategy.
• Conducted simulations comparing DER with existing schedulers (PSLS, SSPPH, FFARS) under varying privacy loads.

Main Results:

• DER demonstrated superior performance across low, medium, and high privacy loads.
• Achieved fewer uncompleted tasks, reduced execution time and cost, and enhanced privacy satisfaction compared to other methods.
• Reference-state-guided initialization and adaptive mutation improved convergence and feasibility preservation.

Conclusions:

• DER effectively manages multi-cloud financial workflows by optimizing multiple objectives simultaneously.
• The proposed methods significantly improve scheduling efficiency and privacy preservation in privacy-aware financial computing.
• DER offers a robust solution for complex multi-cloud scheduling challenges.