RESOURCE PRICING MODEL BASED ON DOUBLE AUCTION FOR THE CLOUDLET FEDERATION

The cloudlet federation is a new computing architecture which reduces the cost of cloudlet deployment and management by sharing cloudlet resources among cloudlet service providers. Renting resources is an important means for the cloudlet federation to improve resource utilization rate. However, the cloudlet federation has the problem of how to formulate a reasonable and effective resource pricing mechanism. Therefore, we construct a resource pricing model based on double auction, aiming at maximizing the benefits of the cloudlet federation. The reasonable dynamic pricing strategies for resource providers and resource demanders are formulated from different perspectives: the asking price of the resource provider considers the resource utilization rate, and the bid of the resource demander considers the task deadline, and the price adjustment strategy is used to conduct multiple rounds of auctions to improve the success rate of task auction. Simulation results show that the proposed dynamic resource pricing algorithm has better performance than the other two fixed resource pricing algorithms in terms of the total benefits of the cloudlet federation, resource utilization rate and auction task success rate.