IEEE Transactions on Industrial Informatics · 2023 · 23 citations · 36 references
ComputingComputer EngineeringComputer ScienceEnergy ManagementSchedulingServerless ComputingCpu SchedulingResource SchedulingResource-constrained SchedulingServerless ArchitectureWorkload SchedulingEnergy ConsumptionEnergy-aware Resource SchedulingEnergy-aware SchedulingFunction SchedulingLinear Programming OptimizationScalabilityScheduling In Serverless ComputingServerless Scheduling
The emerging serverless computing paradigm has recently attracted huge attention from both academia and industry. It brings benefits, such as less operational complexity, high scalability and availability, and lower costs. Serverless applications are usually partitioned into several chains of functions. The serverless provider should schedule functions for execution per customers' requests considering their chained nature. Also, the existing scheduling mechanisms for serverless platforms pay little attention to the reduction of energy consumption during functions' execution. To fill this gap, we present an energy-aware execution scheduler for serverless service providers named EneX. To do so, we formulate the minimization of energy consumption for executing the incoming chains of functions with specified computational loads and deadlines. Due to the intractability of the problem, we introduce a linear programming reformulation based on which, we propose an online scheduler. Finally, our experiments demonstrate the significant improvement of EneX in terms of energy efficiency.
36
The rise of serverless computing
Paul Castro, Vatche Ishakian, Vinod Muthusamy et al. · Communications of the ACM · 2019 · 336 citations