Fault-Aware Service Scheduling Optimization Framework in Edge Data Center

Edge data centers can provide abundant computing, memory, and storage resources for mobile terminal users with low latency to meet the needs of various services and computing tasks. However, due to the uncertainty of its environment compared with the cloud data center, the failure frequency is higher on the edge side, which brings great challenges to user service satisfaction. To mitigate this challenge, minimize the impact of failures on edge data centers, and provide higher reliability for continuous service operation, this paper proposes a fault-aware service scheduling optimization framework. We first explore the impact of the failure scenario and define the state of edge data center resources. Secondly, for the case containing faults, we define it as the target model between fault awareness and task energy efficiency. Finally, a decision-generation technique based on agent learning is proposed to find the service scheduling decision that can minimize the impact of failures. Extensive experiments are conducted, and evaluation results demonstrate that our proposed fault-aware services scheduling technique can effectively improve system reliability and availability.