Dynamic Load Balancing Technology for Cloud-oriented CDN
- School of Computer Science and Technology
150001 Harbin Institute of Technology, Harbin, HL, China
{hehui, wzzhang}@hit.edu.cn - HLJ Province Electronic Information Products Supervision Inspection Institute
150001 Harbin, HL, China
105104397@qq.com, lzg@nstc.gov.cn - Department of Computer Science, University of Houston
Houston, TX, USA
cheng@cs.uh.edu
Abstract
With soaring demands of Internet content services, content delivery network (CDN), one of the most effective content acceleration techniques, is applied into Internet services. Content routing functions in CDN are generally realized by load balancing system. Effectiveness of load balancing strategy determines response speed to users and user experience (UE) directly. This paper extracted the most important influencing factor of CDN loading from common network services and proposed the Variable Factor Weighted Least Connection. The proposed algorithm made real-time computing and dynamic regulation in considering of effect of network applications on server load index, performance changes of the server and workload changes. It has been applied in LVS kernel system successfully. The experiment confirmed that the CDN load schedule system with Variable-Factor-Weighted-Least-Connection could balance loads among cluster servers dynamically according to processing capacity changes of servers, thus enabling to provide users desired services and contents during high large-scale concurrence accesses of users.
Key words
content delivery network (CDN), dynamic load balancing, LVS, Weighted Least Connection
Digital Object Identifier (DOI)
https://doi.org/10.2298/CSIS141104025H
Publication information
Volume 12, Issue 2 (June 2015)
Year of Publication: 2015
ISSN: 2406-1018 (Online)
Publisher: ComSIS Consortium
Full text
Available in PDF
Portable Document Format
How to cite
He, H., Feng, Y., Li, Z., Zhu, Z., Zhang, W., Cheng, A.: Dynamic Load Balancing Technology for Cloud-oriented CDN. Computer Science and Information Systems, Vol. 12, No. 2, 765–786. (2015), https://doi.org/10.2298/CSIS141104025H