A Strategy to Optimize Resource Allocation in Auction-Based Cloud Markets

Current markets of cloud resources are mainly based on a fixed pricing model. In the recent literature several market-based resource allocation models and algorithms have been proposed, showing that dynamic pricing models could result more profitable for both providers and consumers. In this paper we propose a market of resources where the demand and the offer of resources is matched in auction-based sales. Specifically, we looked at this market from the perspective of the provider, who needs a strategy to allocate at best their unused computing capacity. We proposed an adaptive strategy that, suitably customized to the provider's business objective, help them to maximize the revenue in the context of procurement auctions. Furthermore, the impact of resource overbooking onto the utilization level of cloud data centers has been analyzed by means of extensive simulations.