Heterogeneous Resource-aware Load Balancing Scheduling Algorithm in Cloud-native Environments
CHEN Qiliang
HU Zhekun
ZENG Haoyang
Abstract：In addressing the performance and resource utilization deficiencies encountered by high-performance computing tasks and intelligent applications running on cloud-native systems,a heterogeneous resource-aware load balancing scheduling algo-rithm for cloud-native environments is proposed.This algorithm predicts the load of high-performance tasks to implement load bal-ancing for batch task scheduling,which not only balances the utilization of system computing resources but also effectively reduces communication latency between Pods and improves the efficiency of parallel task execution.For intelligent computing tasks requir-ing specific heterogeneous resources,resource reservation is conducted on nodes with idle heterogeneous resources to prevent these tasks from waiting due to CPU resource scarcity.Experimental results show that,compared to the default Kubernetes scheduling al-gorithm,this proposed algorithm reduces the overall runtime of a mixed workload of high-performance and intelligent computing tasks by 8.38%.
Keywords：cloud computingKubernetes schedulingbatch schedulingload balancing
Publication Date：2025-11-20
Online Publishing Date：2026-01-28(First online date of this platform, not the publication date of the document)
Pages：6( 50-55 )