Trading Web Services in a Double Auction-Based Cloud Platform: A Game Theoretic Analysis

Abstract

In the cloud platform, there exist multiple providers offering the same web service, and multiple consumers requiring the same web service. Web service providers (consumers) compete against each other to make a deal with service consumers (providers). Each service provider needs to determine a proper ask to sell the service, while each consumer needs to determine a proper bid to purchase the service. In this paper, in order to address this problem, we assume that the cloud platform runs a double auction marketplace, where web service is traded as the commodity between multiple service providers and consumers. We then use game theory to analyse the Bayes-Nash equilibrium bidding strategies of service providers and consumers on the web service. We also investigate how different service pricing policies adopted by the cloud platform can affect service providers and consumers' Bayes-Nash equilibrium bidding strategies. We show that when the cloud platform sets the uniform market equilibrium price as the service price for all matched service providers and consumers, they will bid close to their valuations on the services. However, when the cloud platform sets discriminatory prices for each matched pair of provider and consumer, service providers will ask more than their valuations and consumers will bid less than their valuations. Moreover, we also investigate the allocative efficiency when the cloud platform adopts different service pricing policies, and show that the cloud platform is very efficient in terms of allocating web services between providers and consumers.