Workflow simulation and multi-threading aware task scheduling for heterogeneous computing

Kelefouras, Vasilios; Djemame, Karim

doi:10.1016/j.jpdc.2022.05.011

Cited by 6 publications

(1 citation statement)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are three general solutions: heuristic-based list scheduling algorithms, random search-based intelligent algorithms, and machine learning-based methods. Some effective heuristic-based list scheduling algorithms have been proposed, such as HEFT (Topcuoglu, Hariri, and Min-You Wu, 2002), PEFT (Arabnejad and Barbosa, 2014), CPOP (Kelefouras and Djemame, 2022). For the random search-based scheduling strategy, microservice scheduling is established as an optimization problem that is solved using intelligent algorithms such as the Genetic Algorithm (Rehman et al, 2019), Ant Colony Algorithm (Gao et al, 2019), and Particle Swarm Algorithm (Rodriguez and Buyya, 2014).…”

Section: Introductionmentioning

confidence: 99%

Service scheduling strategy for microservice and heterogeneous multi-cores-based edge computing apparatus in smart girds with high renewable energy penetration

Hu,

Qu,

Cai

et al. 2024

Front. Energy Res.

View full text Add to dashboard Cite

The microservice-based smart grid service (SGS) organization and the heterogeneous multi-cores-based computing resource supply are the development direction of edge computing in smart grid with high penetration of renewable energy sources and high market-oriented. However, their application also challenges the service schedule for edge computing apparatus (ECA), the physical carrier of edge computing. In the traditional scheduling strategy of SGS, an SGS usually corresponds to an independent application or component, and the heterogeneous multi-core computing environment is also not considered, making it difficult to cope with the above challenges. In this paper, we propose an SGS scheduling strategy for the ECA. Specifically, we first present an SGS scheduling framework of ECA and give the essential element of meeting SGS scheduling. Then, considering the deadline and importance attributes of the SGS, a microservice scheduling prioritizing module is proposed. On this basis, the inset-based method is used to allocate the microservice task to the heterogeneous multi-cores to utilize computing resources and reduce the service response time efficiently. Furthermore, we design the scheduling unit dividing module to balance the delay requirement between the service with early arrival time and the service with high importance in high concurrency scenarios. An emergency mechanism (EM) is also presented for the timely completion of urgent SGSs. Finally, the effectiveness of the proposed service scheduling strategy is verified in a typical SGS scenario in the smart distribution transformer area.

show abstract