Microservice Workflow Scheduling with a Resource Configuration Model Under Deadline and Reliability Constraints.

Area of Science:

• Cloud Computing
• Distributed Systems
• Artificial Intelligence

Background:

• Microservice architectures and containerization are rapidly evolving, increasing the complexity of large-scale cloud service scheduling.
• Ensuring reliability and meeting deadlines for cloud services under these evolving technologies presents a significant challenge.

Purpose of the Study:

• To present a cost-optimized scheduling approach for microservice workflows in container environments.
• To address deadline and reliability constraints in cloud service scheduling.
• To reduce cloud resource leasing costs while maintaining service quality.

Main Methods:

• Developed a graph deep learning model (DeepMCC) for automatic container configuration to meet Quality of Service (QoS) requirements.
• Proposed a reliability microservice workflow scheduling algorithm (RMWS) using heuristic leasing and deployment strategies.
• Evaluated the approach on four scientific workflow datasets.

Main Results:

• The proposed approach achieved an average cost reduction of 44.59% compared to existing reliability scheduling algorithms.
• Demonstrated significant cost improvements, with a 26.63% reduction in the worst case and 73.72% in the best case.
• Successfully met deadline and reliability constraints while optimizing resource utilization.

Conclusions:

• The presented cost-optimized scheduling approach effectively balances reliability, deadline adherence, and cost reduction for microservice workflows.
• DeepMCC and RMWS offer a robust solution for efficient cloud service scheduling in containerized environments.
• The findings highlight the potential for substantial cost savings in cloud resource management through intelligent scheduling.