Multi-Resource Continuous Allocation Model for Cloud Services

Abstract

Cloud service providers deliver a discrete combination of computer resources on demand with pay-as-you-go pricing. However, the discrete combination of computer resources complicates the pricing model, and users often face difficulties in comparing, identifying, and finally selecting a cloud service provider. In an effort to simplify the ever-increasing level of pricing complexity, we propose the continuous allocation model. In this model, a cloud service provider advertises continuous resource sizes while preserving discrete resource allocation. Quantization is used to map between the infinite advertised request sizes and the small number of physical quantized sizes. We use simulation to validate the multi-resource continuous allocation model for cloud services.