ترجمه مقاله نقش ضروری ارتباطات 6G با چشم انداز صنعت 4.0
- مبلغ: ۸۶,۰۰۰ تومان
ترجمه مقاله پایداری توسعه شهری، تعدیل ساختار صنعتی و کارایی کاربری زمین
- مبلغ: ۹۱,۰۰۰ تومان
In cloud computing, the allocation of resources plays a key role in determining the performance, resource utilization and power consumption of the data center. The appropriate allocation of virtual machines in cloud data centers is also one of the important optimization problems in cloud computing, especially when the cloud infrastructure is made of lightweight computing devices. In this paper, we represent the resources’ allocation problem in cloud computing environment as a linear programming model and propose a Hungarian Algorithm Based Binding Policy(HABBP) as a solution for optimizing the model. Finally, we propose an HABBP software implementation as contributed code to the popular CloudSim simulator and compared the HABBP performance to the conventional CloudSim binding policy and a binding based on the Simplex algorithm. Our simulation results show that the newly proposed policy outperforms the conventional binding policy implemented in the CloudSim in terms of jobs total execution time.
I. INTRODUCTION
Cloud computing is a new computing paradigm that provides on-demand network access to large pool of compute, storage and networking resources over the internet [15]. This paradigm offers cost savings because it is based on a pay-asyou-use model where customers pay for computing resources from a service provider only when needed and for only what they use. Cloud computing has three deployment models, namely: public cloud, private cloud, and hybrid cloud [15][16]. Furthermore, Cloud computing services are classified into three types as Infrastructure-as-a-service (IaaS), Platform- asa-Service (PaaS) and Software-as-a-Service (SaaS). IaaS refers to providing hardware infrastructure such as CPU, memory and storage as a service. PaaS refers to providing platforms such as software development frameworks, operating systems or multi-tenant application supports as a service while in SaaS software and applications are provided as a service. These services are provided to the users through virtualization and distributed computing concepts. Virtualization is a technique by which physical infrastructures are abstracted to provide virtualized resources called virtual machines (VMs) for high-level applications [18]. VMs can be homogeneous or heterogeneous.
VI. CONCLUSION AND FUTURE WORK
In this paper, we have revisited the issue of resource allocation in cloud computing environments and proposed HABBP binding policy as a new and optimal solution to the virtual resources allocation problem. HABBP uses a load balancing policy for binding cloudlets to virtual machines in such a way that each cloudlet is allocated to the appropriate virtual machine to optimize the total execution time of completing on-demand user requests. We have formulated an optimization model, proposed HABBP, simulated both HABBP and conventional binding policy in CloudSim and benchmarked these solutions against the Simplex algorithm
6. نتیجه گیری و کار آتی
در این مقاله مساله اختصاص منابع در محیط های رایانش ابری مورد بررسی قرار گرفته است و سیاست ترکیبی HABBP پیشنهادی به عنوان راهکاری جدید و بهینه برای مساله تخصیص منابع مجازی معرفی شده است. الگوریتم HABBP از سیاست توازن بار برای ترکیب تکه ابرها با ماشین های مجازی به نحوی استفاده می کند که هر تکه ابر به ماشین مجازی مناسب اختصاص داده می شود تا کل زمان اجرای درخواست های مورد تقاضای کاربر بهینه شود. ما یک مدل بهینه سازی را فرموله کرده ایم، الگوریتم HABBP را پیشنهاد داده ایم، الگوریتم های HABBP و سیاست ترکیبی متداول را شبیه سازی کرده ایم و این الگوریتم ها را با الگوریتم Simplex مقایسه کرده ایم.