Continuous memory allocation model for cloud services

Abstract

Cloud computing has been growing at a staggering rate by offering a flexible and financially attractive venue for businesses and consumers. Within this context, memory allocation has a significant bearing on cloud-based services. Currently, all major cloud service providers support a small set of discrete memory sizes. We propose a cloud computing service that advertises continuous (any) memory request size, while actually supporting a small number of quantized memory sizes. This scheme redefines and transforms the manner in which cloud services are offered to the public by simplifying the ever-increasing level of pricing complexity. Our proposal targets the root causes of complexity. A service provider with a continuous service model will have a distinct advantage over the competition. We utilize mathematical algorithms to quantize and map the continuous (any) memory request sizes into a small number of quantized sizes optimally, with minimal loss. Furthermore, we investigate different factors affecting the continuous model, such as worth structure, size request distribution, total memory size, and granularity. A simulation is used to conduct our study and confirm our findings.