A Differentiated Service Scheme to Optimize Website Revenues

Publication Type

Journal Article

Publication Date

11-2006

Abstract

We consider a website host server with web quality of service (QoS) capabilities to offer differentiated services. A quantitative modelling framework is set up to analyse the economic benefits of differentiated services and to build optimization models for managing the website host's connection bandwidth to the Internet (which is assumed to be the bottleneck factor determining the QoS). Three models are formulated corresponding to three operational scenarios to provide differentiated services. The first is for the marketing manager to classify visit requests as premium or basic when the information technology (IT) manager has already reserved bandwidths for the two classes, and the second is for the IT manager to allocate the total available bandwidth to each class when the marketing manager has already designated which visit requests are premium and which are basic. The third is for the joint optimization of request classification and bandwidth allocation when centralized coordination is possible. Analytic results are obtained for a special case that corresponds to very impatient customers requesting large amounts of data. Qualitative insights gained and numerical results obtained strongly support the implementation of differentiated services. More interestingly, the decentralized models that use simple and rough-cut rules yield solutions almost as good as the joint optimization model.

Keywords

OR modelling, queueing theory, non-convex optimization, bandwidth allocation, quality of service

Discipline

Operations and Supply Chain Management

Research Areas

Operations Management

Publication

Journal of the Operational Research Society

Volume

57

Issue

11

First Page

1323

Last Page

1340

ISSN

0160-5682

Identifier

10.1057/palgrave.jors.2602100

Publisher

Palgrave

Additional URL

https://doi.org/10.1057/palgrave.jors.2602100

Link to Full Text

Share

COinS