A Framework for Performance Analysis of Computing Clouds

IJSRD Found good research work on Computer Engineering Reseach area


Author :Mintu M. Ladani , Vinit kumar Gupta

College Name :Hasmukh Goswami College of Engineering,Vahelal

Abstract-- Cloud Computing provides data storage capacity and use of Cloud Computing have increased scalability, availability, security and simplicity. As more use of cloud computing environments increases, it is more difficult to deal with the performance of this environments. We have presented Some virtualization and network related communication issues and finally we have designed and implemented modified load balancing algorithm for performance increase. In market use of cloud many issues occurred like as security, privacy, reliability, legal issues, open standard, compliance. so, we have stated to solve these issues such algorithm to assess increase performance of computing clouds. Secondly, i.e. ‘Modified Weighted Active Monitoring Load Balancing Algorithm’ on cloud, for the balancer on Cloud Controller to effectively balance load requests between the available Node Controller, in order to achieve better performance parameters such as load on server and current performance on the server. By Existing Algorithm like in RRA (Round Robin Algorithm) load balance sequentially, we have designed this proposed algorithm on cloud and how to balance load randomly and display by existing algorithm and proposed algorithm comparison. 

Keywords-- Cloud Computing, Modified Weighted Active Monitoring Load Balancing Algorithm, RRA (Round Robin Algorithm)

I. INTRODUCTION 

Framework, users can assess the overhead of acquiring and Releasing the virtual computing resources, they can Different Configurations and they can evaluate different scheduling algorithms. 

II. LITERATURE REVIEW

 Modeling techniques is for workloads and storage devices in order to do load balancing of workloads over a set of devices. The workloads were modeled independent of underlying device using the parameters inherent to a workload such as seek distances, read-write ratio, average IO size and outstanding IOs. For device modeling we used statistics such as IO latency with respect to outstanding IOs which is dependent on the storage device. It is Also described a load balancing engine that can do migrations in order to balance overall load on devices in proportion to their capabilities [18].

•  Research problem: 

1. How to characterize dynamic workloads for load balancing? Are percentile values for workload parameters good enough in real systems?
2. Can we use static values of constants such as K1 to K4 for workloads running on different devices or do we need online estimation?
3. How to measure and predict interference among various workloads accessing a device? 
4. How to suggest storage configuration changes to an administrator based on online workload monitoring. Ultimately the task of workload monitoring, device modeling and load balancing needs to happen in a feedback loop over time to handle churn in today’s storage environments.
5. Cloud computing has emerged as a new technology that lets users deploy their applications in an environment with a promise of good scalability, availability, and fault tolerance. As the use of cloud computing environments increases, it becomes crucial to understand the performance of these environments in order to facilitate the decision to adopt this new technology, and to understand and resolve any performance problems that may appear. In this paper, we present Framework, which is a framework for generating and submitting test overloads to computing clouds. By using A systematic literature review is a means of Identifying, Evaluating and interpreting all available research relevant to a particular research question, or topic area, or phenomena of interest (Kitchen ham, 2004).[14] 

• Reasons for performing SLR 

For full length article and more information click here

Visit us:www.ijsrd.com

like us : www.facebook.com/ijsrd