An Enhanced Load balancing model on cloud partitioning for public cloud Agidi.Vishnu vardhan*1, B.Aruna Kumari*2, G.Kiran Kumar*3 M.Tech Scholar, Dept of CSE, MLR Institute of Technology, Dundigal, Dt: R.R District, Telangana State, India,vishnuvardhan445@gmail.com Associate Professor, Dept of CSE, MLR Institute of Technology, Dundigal, Dt: R.R District, Telangana State, India, aru_bommu@yahoo.com Associate Professor & HOD, Dept of CSE, Marri Laxman Reddy Institute of Technology, Dt: R.R District, Telangana State, India ABSTRACT: Cloud computing is a new and most demanding technology for mutual growth of business environment and software environment. Cloud computing is a distributed computing techniques with huge volume of resources like infrastructure, platform and Applications/software s. An efficient way of using resources in one of the critical task of cloud computing. Load balancing is the core module to distribute the jobs to different servers as well as reducing the load level on each server. We are proposing new model called cloud partitioning for public clouds which will improves the performance all nodes. We are proposing a geographic based method to partition the cloud servers as well as set the refresh period using server statics. KEYWORDS: cloud computing, load balancing model, cloud partitioning, resource scheduling. I.INTRODUCTION Cloud computing is one of the ruling IT technology to make any ones business vision into reality without taking more technical support from different industries. Cloud computing provides a common platform to develop, deploy and host new range of applications by using the help of internet. Cloud computing can be defined in terms cloud and computing. Cloud is a collection of heterogeneous resources connected together to achieve the common goal. It provides an infrastructure for providing services to the end users. Computing specifies set rules and regulations to provide services to the end user by using these resources. Cloud computing can be classified into four types called Public, private, hybrid and community. Public clouds are owned and maintained by the group of companies and provides services to all end users with free of cost or pay-per-use basis, software as a service is belongs to public cloud. Private clouds is a self-service interface that controls common services, allowing IT staff IJCSIET-ISSUE4-VOLUME3-SERIES1 Page 1
to quickly provision, allocate and deliver ondemand IT resources. Highly automated management of resource pools for everything from compute capability to storage, analytics, and middleware. Hybrid Clouds are a composition of two or more clouds (private, community or public) that remain unique entities but are bound together offering the advantages of multiple deployment models. In a hybrid cloud, you can leverage third party cloud providers in either a full or partial manner; increasing the flexibility of computing. Augmenting a traditional private cloud with the resources of a public cloud can be used to manage any unexpected surges in workload. Community A community cloud is a is a multi-tenant cloud service model that is shared among several or organizations and that is governed, managed and secured commonly by all the participating organizations or a third party managed service provider. II.RELATED WORK Load balancing is one of the basic technique through which the set of jobs arrived to cloud can be distributed to the group of servers. A good load balancing model is more general, stable, and scalable which provides less overhead on the system. The main aim of the load balancing model is to improve the performance of systems by balancing loads among them. The load balancing models can be categorized into static which uses the statistical information of the systems or dynamic which considers the current state of the system information. These load balancing models can be either cooperative or non cooperative. There are many load balancing algorithms proposed to improve the overall performance of the system. Some of the good algorithms are Round Robin, Equally spread current Execution Algorithm, and Ant Colony algorithms. Tejinder Sharma, Vijay Kumar Banga [2] proposed an algorithm which considers all the systems should have equal load at any time. Nishnt et al. [3] used the ant colony optimization method in nodes load balancing. Randles et al [4] gave a compared analysis of different load balancing algorithms by considering the performance and cost. They finalized that ESCE algorithm and threshold algorithm are better than round robin algorithm. Some of the classical algorithms are similar to operating system algorithms like FCFS. III.PROPOSED METHODOLOGY The proposed system is mainly concentrates on the public cloud. The Public cloud contains many nodes which are located in many areas. The main of our proposed system is partition the existed cloud into different groups based on the geographical information of the nodes. Assigning jobs to the nodes based on the load and geographical status. The basic two modules are 3.1 partitioning of public cloud 3.2 Job allocation to the nodes IJCSIET-ISSUE4-VOLUME3-SERIES1 Page 2
3.1 PARTITIONING OF PUBLIC CLOUD Partitioning of cloud uses geographical information as the base. If the number of nodes in cloud increased in the same region then the nodes are partitioned into two parts with a common load balancer. This balancer will maintain the statistical information of nodes like current state, load, available CPU time etc. 3.2 JOB ALLOCATION TO NODES Job allocation to the nodes is very important task in our proposed system which improves the overall performance of the cloud. The basic job allocation includes two major components called main controller and balancer. The main controller is job is to collect the jobs and find less load balancers in the corresponding regions of the jobs. The balancer checks the load on the each node and allocates the jobs to the less load nodes. Here the balancer can have two local partitions. The partition with good efficiency and fewer loads will be selected and then job is assigned to the particular node. Figure 1: Basic Architecture of load balancer Job allocation considers different states of the nodes. Basing those nodes status the jobs are allocated. The main states of the nodes are Idle, Normal and Overload. Idle state is nothing but the nodes are free and ready to execute the tasks. Normal state means even the node executing some task still available for further jobs executions. Overloaded is a critical state of the node which cannot execute any jobs right now. The following flow chart explains how the jobs allocating. IJCSIET-ISSUE4-VOLUME3-SERIES1 Page 3
IV.CONCLUSION AND FUTURE ENHANCEMENT Load balancing model will improves overall performance of the systems. The proposed system is an efficient one to allocate the jobs to nodes and monitor and manage all the nodes to be efficient to execute any jobs. Public cloud partition will reduces the overhead on balancers and improves the nodes performance. V.REFERENCES Figure 2: Flow chart of job allocation. The final and important point is to refresh the main controller and balancers. This can be done on the previous statistics of the nodes. If the nodes are taking more time to execute the previous jobs the refresh time is increased else it will be decreased. [1] Gaochao Xu, Junjie Pang, and Xiaodong Fu A Load Balancing Model Based on Cloud Partitioning for the Public Cloud IEEE TRANSACTIONS ON CLOUD COMPUTING YEAR 2013 [2] Tejinder Sharma, Vijay Kumar Banga Efficient and Enhanced Algorithm in Cloud Computing International Journal of Soft Computing and Engineering (IJSCE) ISSN: 2231-2307, Volume-3, Issue-1, March 2013 [3] K. Nishant, P. Sharma, V. Krishna, C. Gupta, K. P. Singh, N. Nitin, and R. Rastogi, Load balancing of nodes in cloud using ant colony optimization, in Proc. 14 th International Conference on Computer Modelling and Simulation (UKSim), Cambridgeshire, United Kingdom, Mar. 2012, pp. 28-30. [4] M. Randles, D. Lamb, and A. Taleb- Bendiab, A comparative study into distributed load balancing algorithms for cloud computing, in Proc. IEEE 24 th International Conference on Advanced Information Networking and Applications, Perth, Australia, 2010, pp. 551-556. [5] Radojevic, B. & Zagar, M. (2011). Analysis of issues with load balancing algorithms in hosted (cloud) environments In proceedings of 34th International Convention on MIPRO, IEEE. IJCSIET-ISSUE4-VOLUME3-SERIES1 Page 4
[6] Bhasker Prasad Rimal, Eummi Choi, Lan Lump (2009) A Taxonomy and Survey of Cloud Computing System, 5th International Joint Conference on INC, IMS and IDC, IEEE Explore 25-27Aug 2009, pp. 44-51 [7] Bhathiya, Wickremasinghe.(2010) Cloud Analyst: A Cloud Sim-based Visual Modeler for Analysing Cloud Computing Environments and Applications [8] C.H.Hsu and J.W.Liu(2010) "Dynamic Load Balancing Algorithms in Homogeneous Distributed System," Proceedings of The 6thInternational Conference on Distributed Computing Systems,, pp. 216-223. [9] Martin Randles, David Lamb, A. Taleb- Bendiab(2010), A Comparative Study into Distributed Load Balancing Algorithms for Cloud Computing IEE 24thInternational Conference on Advanced Information Networking and Applications Workshops. [10] Martin Randles, David Lamb, A. Taleb- Bendiab(2010), A Comparative Study into Distributed Load Balancing Algorithms for Cloud Computing, IEEE 24th International Conference on Advanced Information Networking and Applications Workshops, 20-23, pp. 551-556. [11] Mishra, Ratan, Jaiswal, Anant,P Ant Colony Optimiza tion: A Solution Of Load Balancing In Cloud, April 2012, International Journal Of Web & Semantic Technology;Apr2012, Vol. 3 Issue 2, P33 [12] Eddy Caron, Luis Rodero-Merino Auto-Scaling, Load Balancing And Monitoring In Commercial And Open- Source Clouds Research Report,January2012. [13] Nidhi Jain Kansal, Cloud Load Balancing Techniques: A Step Towards Green Computing, IJCSI International Journal Of Computer Science Issues, January 2012, Vol. 9, Issue 1, No 1,, Pg No.:238-246, ISSN (Online): 1694-0814. IJCSIET-ISSUE4-VOLUME3-SERIES1 Page 5