Service value optimization of cloud hosted systems using particle swarm technique

Cloud Computing has become a more promising technology with potential opportunities, through reducing the high cost of running the traditional business applications and by leading to new business models. Nonetheless, this technology is fraught with many challenges. From a Software as a Service (SaaS) provider perspective, deployment choices are one of the major perplexing issues in determining the degree to which the application owners’ objectives are met while considering their customers’ targets. The purpose of this paper is to present a new model that allows the service owner to optimize the resources selection based on defined metrics when responding to many customers’ with various priorities.

More than 65 academic papers have been collected, a short list of the most related 35 papers have been reviewed, in addition to assessing the functionality of major cloud systems. A potential set of techniques has been investigated to determine the most appropriate ones. Moreover, a new model has been built and a study of different simulation platforms has been conducted.

The findings demonstrate that serving many SaaS customer requests, with different agreements and expected outcomes, would have mutual influence that impact the overall provider objectives. Furthermore, this paper investigates how tagging those customers with various priorities, with reflection of their importance to the provider, permits controlling and aligning the selection of computing resources as per the current objectives and defined priorities.

This study provides researchers with a useful literature, which can assist them in relevant subject. Additionally, it uses a value-based approach and particle swarm technique to model and solve the optimization of the computing resource selection, considering different business objectives for both stakeholders, providers and customers. This study derives priority of a number of factors, by which service providers can make strong and adaptive decisions.

The paper includes implications on how the SaaS service provider can make decisions to select the needed virtual machines type driven by his own preferences.

This paper rests on the usage of Particle Swarm Optimization technique to optimize the business value of the service provider, as well as the usage of value-based approach. This will help model that value in order to combine the total profit of the provider and the customer satisfaction, based on the agreed budget and processing time requested by the customer. Another additional approach has been charted by using the customer severity factor that allows the provider to reflect the customer importance while making the placement decision.