On the effect of using third-party clouds for maximizing profit

Services in cloud computing systems are typically categorized into three types-software as a service (SaaS), platform as a service (PaaS) and infrastructure as a service (IaaS). These services can be prepared in the form of virtual machine (VM) images; and they can be deployed and run dynamically as clients request. Since the cloud service provider has to deal with a diverse set of clients, including both regular and new/one-off clients, and their requests most likely differ from one another, the judicious scheduling of these requests plays a key role in the efficient use of resources for the provider to maximize its profit. In this paper, we address the problem of scheduling arbitrary service requests of those three different types-taking into account the maximization of profit-in cloud environments, and present the client satisfaction oriented scheduling (CSoS) algorithm. Our algorithm effectively exploits different characteristics of those three service types and the availability of third-party cloud service providers who have (or are capable of having) identical service offerings (using virtual machine images). Our main contribution is the incorporation of client satisfaction into our request scheduling; this incorporation enables to increase profit by avoiding the discontinuation of service requests from those unsatisfied clients due to the poor quality of service.