.png){kind=icon}
返回
DoME: Dew computing based microservice execution in mobile edge using Q-learning
.png){kind=icon}
PRE
AI
DOI：10.1007/s10489-022-04087-x.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
The microservice approach unlatched a new window of distributed service-oriented architecture over the computational horizon. Microservice coordination in a heterogeneous and distributed manner among different edge computing nodes with minimum delay is the challenge of interest. Service migration to vantage locations accompanying bulk data hinge on bandwidth and internet connectivity. This paper proposes a Dew Computing-based Microservice Execution (DoME) scheme using reinforcement learning to alleviate the comprehensive service delay and provide seamless and real-time responses with optimized costs. With internet connectivity, DoME elects the edge server for service migration in exigency. During deficient internet connection, the service execution comes to pass at the dew server of the novel dew-cloud framework. This study imitates the Q-learning method comprehending the dew-edge-cloud framework and develops the Dew-Q algorithm to execute microservices on the move without continuous internet connectivity. The Dew-Q algorithm uses an agent that intelligently assimilates from historical data to envisage the service migration-node selection with the uphold connectivity. The proposed scheme is compared with existing microservice execution policies like FWS and SDGA schemes. The performance evaluations exhibit that the DoME scheme excels 41% to 68% compared with FWS and 12% to 86% with SDGA in terms of execution and migration cost. The proposed scheme dynamically controls microservice execution from the perspective of migration and timeliness with intermittent internet connectivity.
Keyword:
Microservice execution
Service migration
Dew computing (DC)
Q-learning
Mobile edge computing (MEC)
期刊
IF:
3.5
论文数:
7.6K
被引数:
1.7W
机构
引用论文
.png){kind=icon}
Correlation analysis of cytokine gene expression changes and antiulcer effects of Pro-Gly-Pro and N-Acetyl-Pro-Gly-Pro in experimental models of ulceration细胞因子基因表达变化与Pro-Gly-Pro和N-Acetyl-Pro-Gly-Pro抗溃疡效应的相关性分析在溃疡实验模型中的应用
PREAI
分享收藏Delay-Aware Microservice Coordination in Mobile Edge Computing: A Reinforcement Learning Approach移动边缘计算中的延迟感知微服务协调: 一种强化学习方法
PREAI
分享收藏分享收藏Service Candidate Identification from Monolithic Systems Based on Execution Traces基于执行轨迹的单片系统候选服务识别
OAAI
OAAI分享收藏Towards the optimality of service instance selection in mobile edge computing移动边缘计算中服务实例选择的最优性研究
PREAI
分享收藏
查看更多内容