Multiworkflow Scheduling in Edge-Cloud Computing by African Vulture Optimization Algorithm

With the ongoing expansion of user demands, the proliferation of application workflows in the cloud has surged. Application service providers concentrate on their fundamental business requirements to enhance service quality. They construct edge-cloud platforms utilizing containers, microservices, an...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2024 11th International Forum on Electrical Engineering and Automation (IFEEA) s. 1426 - 1432
Hlavní autoři: Wang, Ze, Zhu, Qinghua, Hou, Yan
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 22.11.2024
Témata:
Carbon dioxide
Cloud computing
Containers
edge-cloud computing
Heuristic algorithms
intelligent optimization algorithms
microservice
Microservice architectures
Optimization
Processor scheduling
Real-time systems
renewable energy
Renewable energy sources
Virtual machines
workflow scheduling
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:With the ongoing expansion of user demands, the proliferation of application workflows in the cloud has surged. Application service providers concentrate on their fundamental business requirements to enhance service quality. They construct edge-cloud platforms utilizing containers, microservices, and other technologies to oversee various applications and judiciously distribute cloud infrastructure resources in response to user demands. Effectively allocating resources in the cloud has emerged as the pivotal challenge for workflow scheduling in the containerized edge-cloud environment. However, the container layer adds complexity to task scheduling in edge-cloud environments due to the auto-scaling of containers. Existing algorithms are inadequate for the two-tier architecture based on virtual machines and containers in an edge-cloud computing environment integrated with renewable energy sources. This study aims to address multiworkflow task scheduling in such a scenario. Firstly, we establish multiple optimization objectives for task scheduling of multiworkflow on microservices in an edge-cloud environment. Secondly, we adopt the reinforcement multiobjective African vultures algorithm (RMOAVOA) to minimize makespan and carbon emissions. This algorithm employs adaptive grid techniques and a migration-based VM selection policy to find the optimal VM task migration when an infeasible solution occurs. Further, an actor-critic framework is used to sort tasks before scheduling. Additionally, we utilize a method for reusing containers to adjust task-container mappings, thereby saving time. Finally, simulation experiments validate the effectiveness of our proposed approach by comparing it with benchmark algorithms.
DOI:10.1109/IFEEA64237.2024.10878706