چكيده به لاتين
Nowadays, cloud computing environment has attracted a lot of attention due to its characteristics such as elasticity in resource allocation and management, flexibility, dynamism, scalability, and multitenancy. Users of cloud computing environments can use computing services in three service models named IaaS, PaaS, and SaaS. Various virtual machine scheduling and placement mechanisms have been introduced by researchers to resource efficiency enhancement, cost reduction, QoS assurance, energy consumption reduction, etc. Combinatorial workloads which contains a combination of both I/O and CPU intensive workloads, have a significant presence in the cloud computing environments. Nevertheless, a few studies have been conducted to measure impact of combinatorial workloads on response time and little efforts have been done to reduce response time by using workload-aware virtual machine placement mechanisms. In this research, a new workload-aware virtual machine scheduling and placement mechanism has been introduced, which aims to reduce response time of virtual machines with combinatorial workloads. Implementation of the proposed virtual machine placement mechanism is confined to the scheduler part of Nova (the part of OpenStack that manages compute resources). Through the experiments on a multi-node private cloud deployed using OpenStack cloud platform and Xen hypervisor, the proposed virtual machine placement mechanism has been shown to reduce the response time of virtual machines with combinatorial workloads.
