Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Online pricing for Web service providers
Esmaeilsabzali S., Day N.  Economics driven software engineering research (Proceedings of the 2006 International Workshop on Economics Driven Software Engineering Research, Shanghai, China, May 27, 2006)37-42.2006.Type:Proceedings
Date Reviewed: Aug 24 2006

In the provisioning of Web services, service providers cannot accept all requests for their services early; to do so would exhaust their capacities early, missing the chance to accept more profitable requests later. Conversely, service providers cannot be too conservative and reject all requests for their services. The focus of this paper is on the problem of the algorithmic complexity of request selection and pricing for Web service providers, when user requests arrive over time. Web services are considered to be pay-per-view functionalities that are invoked over the Internet.

A theoretical analysis is done to determine which variants of this problem are algorithmically nondeterministic polynomial time (NP) complete. From a practical point of view, many of the assumptions made in this paper do not hold: first, the service level agreements, billing bundles, and the service provider’s control over provisioning delays restrict the complexity. On the other hand, peering, needed transmission capacity adaptation, and content updates all add to the complexity.

The paper considers uniform and dynamic pricing. For uniform pricing, it is stated that there exists no online algorithm with a constant competitive ratio. For dynamic pricing, in its general form, it is stated that there does not exist any known online algorithm with a constant competitive ratio. For the complexity determination, the online version of dynamic pricing is compared with the dynamic (online) knapsack problem, where items arrive according to a Poisson process in time. This assumption is quite surprising, as it is almost absent from Internet traffic. A third analysis deals with a randomized, threat-based, online algorithm that tries to maintain a certain level of competitiveness in comparison to the maximum possible profit.

The paper offers an attempt to determine the algorithmic complexity of some theoretical instance-based pricing schemes, under some assumptions not found in real Web services (whether broadband or wireless). The consequences of the result are not spelled out.

Reviewer:  Prof. L.-F. Pau, CBS Review #: CR133225 (0709-0907)
Bookmark and Share
  Reviewer Selected
 
 
Online Computation (F.1.2 ... )
 
 
Economics (K.6.0 ... )
 
 
Communications Applications (H.4.3 )
 
Would you recommend this review?
yes
no
Other reviews under "Online Computation": Date
Online computation and competitive analysis
Borodin A., El-Yaniv R., Cambridge University Press, New York, NY, 1998. Type: Book (9780521563925)
Mar 1 1999
A primal-dual randomized algorithm for weighted paging
Bansal N., Buchbinder N., Naor J. Journal of the ACM 59(4): 1-24, 2012. Type: Article
Feb 1 2013
An introduction to online computation: determinism, randomization, advice
Komm D., Springer International Publishing, New York, NY, 2016.  349, Type: Book (978-3-319427-47-8)
Oct 3 2017

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy