Resource management plays a key role in the cloud-computing environment in which applications face with dynamically changing workloads. However, such dynamic and unpredictable workloads can lead to performance degradation of applications, especially when demands for resources are increased. To meet Quality of Service (QoS) requirements based on Service Level Agreements (SLA), resource management strategies must be taken into account. The question addressed in this research includes how to reduce the number of SLA violations based on the optimization of resources allocated to users applying an autonomous control cycle and a fuzzy knowledge management system. In this paper, an adaptive and fuzzy resource management framework (AFRM) is proposed in which the last resource values of each virtual machine are gathered through the environment sensors and are sent to a fuzzy controller. Then, AFRM analyzes the received information to make decision on how to reallocate the resources in each iteration of a self-adaptive control cycle. All the membership functions and rules are dynamically updated based on workload changes to satisfy QoS requirements. Two sets of experiments were conducted on the storage resource to examine AFRM in comparison to rule-based and static-fuzzy approaches in terms of RAE, utility, number of SLA violations, and cost applying HIGH, MEDIUM, MEDIUM-HIGH, and LOW workloads. The results reveal that AFRM outweighs the rule-based and static-fuzzy approaches from several aspects.

中文翻译：

---

云计算中一种自适应模糊资源管理方法

资源管理在应用程序面临动态变化的工作负载的云计算环境中起着关键作用。但是，这种动态且不可预测的工作负载会导致应用程序性能下降，尤其是在对资源的需求增加时。为了满足基于服务水平协议 (SLA) 的服务质量 (QoS) 要求，必须考虑资源管理策略。本研究解决的问题包括如何基于分配给用户的资源的优化应用自主控制循环和模糊知识管理系统来减少 SLA 违规次数。在本文中，提出了一种自适应和模糊资源管理框架（AFRM），其中通过环境传感器收集每个虚拟机的最后资源值并将其发送到模糊控制器。然后，AFRM 分析接收到的信息以决定如何在自适应控制循环的每次迭代中重新分配资源。所有成员函数和规则都根据工作负载变化动态更新以满足 QoS 要求。在存储资源上进行了两组实验，以在 RAE、效用、SLA 违规次数和应用高、中、中高和低工作负载的成本方面与基于规则的和静态模糊方法进行比较来检查 AFRM . 结果表明，AFRM 从几个方面优于基于规则和静态模糊的方法。