Dynamic Balancing of Complex Event and Data Recovery in Cloud Environment Pranjali Dhore PG Student, Department of Information Technology, MITCOE, Pune, India dhore.pranjali08@gmail.com Kishor Kolhe Associate Professor, Department of Information Technology MITCOE, Pune, India krkolhe@gmail.com Abstract- Distributed computing is a developing innovation deals with dispersed figuring system. . One of the most challenging issues in cloud Computing is efficient scheduling of tasks. Cloud computing offers a variety of dynamic load balancing methods. Load balancing is the methodology of distributing the load among different node of a distributed framework to enhance both resource usage and reaction time while likewise keeping away from a circumstance where a percentage of the node are intensely stacked while different node are sit out of gear or doing next to no work. An solution for unbalance circumstance is to utilize parallelization approaches yet at the same time node will stay overwhelming. In this paper, we propose an integrated Dynamic Load balancing algorithm to attain scalability even in heavy queries, avoid fault tolerance when system crash, event generation and handling, job submission control, Overload control on cloud environment. Keywords— Cloud Computing, Complex Event processing, Distributed database, Dynamic Load balancing, Scalability. I. INTRODUCTION Now days, distributed database is spread over the system where application is required. High rate transforming of information is most urgent in appropriated environment.. Complex Event Processing is event processing that combines data from multiple sources. With the help of this event processing system thousands of incoming queries can be register, processed and that will produces result periodically, this will lead to scalability, failure of node because of heavily loaded node problem when big database is used. Fig 1: Components of Cloud Computing Solution Fault tolerance occurred because of load unbalancing condition means a situation where some nodes are heavily loaded and some nodes are idle. To avoid fault tolerance nodes, the method is required to allocate dynamic workloads equally to all the nodes across cloud network hence load balancing approach is used. Load balancing is the method which makes sure that every processor within the system or every node in the network consume equal amount of power and finish approximately equal amount of work at any instant of time. The load can be identified as data uploading capacity, CPU load or network delay. In this paper, we proposed a robust architecture which schedules the queries for scalability and distribution of equal work load. Registered incoming CEP queries in system are transferred to idle nodes or not heavily node. To improve both resource utilization and performance scheduler is responsible to distribute the load among various nodes and make it balance. Queries are applied to scheduler module of system, then scheduler schedules the query to lightly weighted node with respect to CPU capacity, memory consumption. Incoming queries are delivered to queue; if queue is full or reaching to limit size in queue measure will be taken. Overload situation can kill the nodes and results into data loss but this will be avoided by using replication method. The proposed system maintain ready queue, use it to queue incoming query when nodes are loaded and none of node is free. At last event is generate if no selection is conceivable. The rest of the paper is organized as follows. Section II discusses literature review, Section III describes the problem statement, and Section IV discusses proposed method. We conclude the paper in last section. II. LITERATURE REVIEW There are numerous methodologies taken inside the writing for learning load balancing. In appropriated systems adjusting the node in a versatile way enhances the system execution impressively. Further, a definitive objective of load balancing is as following [5]  Even appropriation of load to every node.  Minimization of handling time for every node.  Maximum utilization of each resource. AVCOE, Sangamner iPGCON-2015 SPPU, Pune 24th & 25th March 2015 Fourth Post Graduate Conference Page 1 of 4