Usage-Oriented Resource Allocation Strategy in Edge Computing Environments

Hsieh, Tsu-Hao; Ho, Kuan-Yu; Tsai, Meng-Yo; Lai, Kuan‐Chou

doi:10.1007/978-3-030-39746-3_37

Cited by 2 publications

(7 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the existing scoring strategies, such as LeastRequestedPriority and BalancedResourceAllocation, mostly use the cluster's idle rate and cluster's balance degree. 34,35 Kubernetes can define auto-scaling strategies, but more for the cloud. 36,37 The container scheduling problem is usually formulated as a multi-objectives optimization problem.…”

Section: Related Workmentioning

confidence: 99%

“…(3) LeastRequestedPriority (LRR) 34 : It is a scoring function provided by Kubernetes, which determines the score by the ratio of idle resources to the total amount of resources. It tries to schedule Pods to nodes that consume relatively small computing resources.…”

Section: Baseline Approachesmentioning

confidence: 99%

“…where U cpu represents the consumed CPU on the node, U mem represents the consumed memory on the node, C cpu represents the total CPU capacity of the node, and C mem represents the total memory capacity of the node. ( 4) BalancedResourceAllocation (BRA) 34 : It is also a scoring function provided by Kubernetes, which is to calculate the difference between CPU utilization and memory utilization. The smaller the difference, the more balanced the CPU and memory utilization, and the higher the score.…”

Section: Baseline Approachesmentioning

confidence: 99%

“…The default container deployment strategy provided by Kubernetes is multi‐criteria decision‐making model, which calculates the score of each node, and the value is score is between 0 and 10. However, the existing scoring strategies, such as LeastRequestedPriority and BalancedResourceAllocation, mostly use the cluster's idle rate and cluster's balance degree 34,35 . Kubernetes can define auto‐scaling strategies, but more for the cloud 36,37 .…”

Section: Background and Related Workmentioning

confidence: 99%

“…In addition, three existing baseline approaches were compared, including LRR, BRA, and OLBP. DICS‐OPT : Data‐intensive container scheduling with optimized weighting factors which are obtained using the genetic algorithm. DICS‐EQ : This approach uses the same weighting factors, and all the five weighting factors are set to 0.2. LeastRequestedPriority (LRR) 34 : It is a scoring function provided by Kubernetes, which determines the score by the ratio of idle resources to the total amount of resources. It tries to schedule Pods to nodes that consume relatively small computing resources.…”

Section: Performance Evaluationmentioning

confidence: 99%

See 4 more Smart Citations

Container‐based data‐intensive application scheduling in hybrid cloud‐edge collaborative environment

2023

View full text Add to dashboard Cite

Container virtualization technology represented by Docker has been widely used in the industry due to its advantages of lightweight, fast deployment, and easy portability. It can bring convenience to system deployment, operation, and maintenance. This article considers the implementation of AI‐based IoT applications based on Docker and container technology in a cloud‐edge collaborative environment. Three typical data‐intensive applications, preprocessing, training, and inference in machine learning, are deployed using Docker containers. For these three types of tasks, this article proposes a container‐based data‐intensive application scheduling framework. Using an attribute‐based data‐driven method, a new scheduling approach considering multiple weighting factors is proposed, which is called DICS‐OPT. The cloud nodes and edge nodes are scored and sorted uniformly. The container is scheduled to the node with the highest score, and then the task data is transmitted to the node to execute the task. NSGA‐III‐based genetic algorithm has been proposed to search for the optimal weighting factors. This article introduces the implementation of the prototype system and builds a cloud‐edge collaborative testbed consisting of Raspberry Pis and PCs. The performance evaluation results indicates that the proposed scheduling approach outperforms existing approaches. Compared with existing approaches, DICS‐OPT improves the average edge resource utilization by 10.05% to 69.04%, and saves the cloud resource cost by 16.02% to 36.68%.

show abstract