A Credit-Based Allocation Method of Resource Quotas for Cloud Users

Abstract

Many studies have been conducted on resource allocation and scheduling methods. While these methods seldom give effective measures to prevent resource abuse. To some extent, the restriction of resource quota relieves the problem of resource abuse in regard to public cloud platform. The common allocation method mainly adopts a fixed resource quota to limit an user's resource usage for the user's first application. When the user requires more resources, he is asked to apply for a new resource quota again. Generally, the administrator of cloud platform simply checks his resource demand and gives him the requested quota. However, in some cases, an user can occupy the quota with low resource utilization in long time, and even he can lie to deceive more resource quota. These may lead a single user to exceedingly occupy resource quota, while other users cannot gain enough resource quota. The fairness in resource quota allocation is violated. To address this issue, we propose a credit-based allocation method of resource quotas for cloud users in this paper. The key idea is to first set up an user's credit factor according to the user's historical resource quota utilizations, and then allocate the user new resource quota based on it. We conduct three simulated experiments on our method. The experimental results show that our method imposes punitive allocation on malicious behaviours, such as occupying resource quota with low quota utilization in long time, deceiving more resource quota and etc. Besides, our method can also satisfy more users' resource quota demands with respect to limited resource. Compared with the existing common method, our method strengthens the fairness in resource quota allocation effectively.