Towards Understanding Cloud Usage through Resource Allocation Analysis on XSEDE

نویسندگان

  • Hyungro Lee
  • Gregor von Laszewski
  • Fugang Wang
  • Geoffrey C. Fox
چکیده

In shared resource environments, usage data is necessary to identify utilization of the infrastructure by users. Many cloud platforms recently started to collect measurements for use of resources that can be applied to billing and monitoring. Understanding utilization and performance through these measurements is crucial in the infrastructure in order to provide better cloud provisioning, system management and capacity planning. In this paper, we present an integrated cloud accounting solution on XSEDE, Cloud Metrics, to measure cloud resource usage across several cloud platforms such as OpenStack, Nimbus, and Eucalyptus. The usage data allows a user to see as how all resources are efficiently supplied to their applications and discover patterns from cumulative data. With Cloud Metrics, virtual resources such as compute, storage and network are measured to evaluate time and cost of user applications and the statistics for these resources offer visibility to utilization of XSEDE cloud resources. This article shows statistical analysis of several case studies by tracing resource allocation on FutureGrid as XSEDE resources. Based on the observation on FutureGrid, we found that different patterns between scientific research projects and educational projects regarding the type of virtual machines (VMs) and the patterns of using virtual resources. Cloud Metrics enables users and project leaders to identify utilization and performance on XSEDE.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Comparing the Consumption of CPU Hours with Scientific Output for the Extreme Science and Engineering Discovery Environment (XSEDE)

This paper presents the results of a study that compares resource usage with publication output using data about the consumption of CPU cycles from the Extreme Science and Engineering Discovery Environment (XSEDE) and resulting scientific publications for 2,691 institutions/teams. Specifically, the datasets comprise a total of 5,374,032,696 central processing unit (CPU) hours run in XSEDE durin...

متن کامل

FRA-PSO: A two-stage Resource Allocation Algorithm in Cloud Computing

Cloud computing gives a large quantity of processing possibilities and heterogeneous resources, meeting the prerequisites of numerous applications at diverse levels. Therefore, resource allocation is vital in cloud computing. Resource allocation is a technique that resources such as CPU, RAM, and disk in cloud data centers are divided among cloud users. The resource utilization, cloud service p...

متن کامل

A review of methods for resource allocation and operational framework in cloud computing

The issue of management and allocation of resources in cloud computing environments, according to the breadth of scale and modern technology implementation, is a complicated issue. Issues such as: the heterogeneity of resources, resource dependencies to each other, the dynamics of the environment, virtualization, workload diversity as well as a wide range of management objectives of cloud servi...

متن کامل

Integrated modeling and solving the resource allocation problem and task scheduling in the cloud computing environment

Cloud computing is considered to be a new service provider technology for users and businesses. However, the cloud environment is facing a number of challenges. Resource allocation in a way that is optimum for users and cloud providers is difficult because of lack of data sharing between them. On the other hand, job scheduling is a basic issue and at the same time a big challenge in reaching hi...

متن کامل

Virtual Machine Customization and Task Mapping Architecture for Efficient Allocation of Cloud Data Center Resources

Energy usage of large-scale data centers has become a major concern for cloud providers. There has been an active effort in techniques for the minimization of the energy consumed in the data centers. However, most approaches lack the analysis and application of real cloud backend traces. In existing approaches, the variation of cloud workloads and its effect on the performance of the solutions ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2014