A feedback-based combinatorial fair economical double auction resource allocation model for cloud computing

Abstract

Resource provisioning by cloud computing is one of the emerging fields for providing computing services anywhere at any time without physical barriers. It not only reduces the cost of set-up and maintenance of infrastructure at the on-site level but also helps in optimal resource utilization and reduction of e-waste generation to save the environment. It provides versatile resources to run different types of applications efficiently without worrying about delay or failure. Auction emerges as one of the most promising pricing schemes for the provisioning of resources in cloud computing. Providers and customers have different mind-set. The providers desire to earn more profit and lead the market, while customers prefer economical rates with high quality of service for the resources. In the virtual world, the providers’ genuineness is crucial, and the feedback from the customers plays a vital role that quantifies this genuineness.This paper proposes a Feedback-based Combinatorial Fair Economical Double Auction Resource Allocation Model (FCFEDARA) that investigates the genuineness of providers based on the offered prices for the required resources along with the feedback given by the customers. The proposed framework provides a common platform for the customers to access resources from different providers at optimal prices, as well as prioritizes genuine providers with good feedback over in-genuine providers with bad feedback. In the combinatorial double auction model, customers and providers submit their bundle bid along with a list of resources in the auction. The proposed model takes care of the truthfulness of the providers, penalizes market spoilers, and gives priority to those providers who have good customer feedback. Three different cases with their implications have analyzed, and evaluated based on the bid, configuration of resources, and customer feedback. Case 1 focuses on feedback given by the customers. In Case 2 and Case 3, providers’ genuineness, as well as their feedback from the customers, are key factors to determine winner providers. It is in contrast with earlier models where genuineness is based only on the bid and configuration of resources offered by providers. The proposed model was simulated on the CloudSim simulator, a Java-based simulator that provides a cloud computing environment, and results are evaluated and compared with two popular existing models: CDARA, its variant model, and the author’s previous approach (CFEDARA). The experimental results and the graphs have clearly shown the promotion of genuine providers with good feedback, fair allocation with economical pricing while maintaining the overall profit of the system. The participation ratio of customers and providers is maintained to sustain demand and supply in the virtual market.