Load balanced task scheduling for cloud computing: a probabilistic approach

Load balancing is the process of distributing customer tasks among multiple computing resources, such as virtual machines (VMs), servers and networks. It is a major concern in cloud computing as the number of customers demanding the service is growing exponentially. An efficient load balancing appro...

Full description

Saved in:
Bibliographic Details
Published in:Knowledge and information systems 2019-12, Vol.61 (3), p.1607-1631
Main Authors: Panda, Sanjaya K., Jana, Prasanta K.
Format: Article
Language:English
Subjects:
Algorithms
Cloud computing
Computer Science
Computer simulation
Confidence intervals
Data Mining and Knowledge Discovery
Database Management
Information Storage and Retrieval
Information Systems and Communication Service
Information Systems Applications (incl.Internet)
IT in Business
Load
Load balancing
Probability theory
Regular Paper
Statistical analysis
Statistical tests
Task scheduling
Variance analysis
Virtual environments
Virtual networks
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Load balancing is the process of distributing customer tasks among multiple computing resources, such as virtual machines (VMs), servers and networks. It is a major concern in cloud computing as the number of customers demanding the service is growing exponentially. An efficient load balancing approach can monitor the load of the VMs proactively and assigns the customer tasks to the VMs accordingly. This paper presents a load balanced task scheduling algorithm in the cloud, which is based on probability theory. The proposed algorithm is shown to be a 2-approximation algorithm with a time complexity of O ( lm ), where l is the number of customer tasks and m is the number of VMs. The algorithm is simulated extensively. The simulation results demonstrate that our proposed algorithm can remarkably balance the load of the VMs as compared to the existing algorithms in four different performance measures, namely standard deviation of VM loads, maximum load, minimum load and zero load. The performance is also validated through statistical test by means of analysis of variance and 95% confidence interval.
ISSN:0219-1377
0219-3116
DOI:10.1007/s10115-019-01327-4