High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

A Novel Approach for Transaction Management in Heterogeneous Distributed Real Time Replicated Database Systems


Mr. Dheeraj Bhimrao Lokhande , Solapur University,Solapur; Prof. Dhainje Prakash B., SIETC, Paniv, Solapur, India


Cloud Computing, Cloud Storage, Workload Characterization, Energy Efficiency, Resource Management, Resource Allocation


Today, Cloud computing is most widely used system due to number of benefits to end users. In IT (Information Technology) organizations, cloud computing is most vital domain to work. There are different types of services such as SaaS, IaaS and PaaS provided by cloud computing based on the end users needs. Rather than using own resources for data storage and Management, organizations started to utilize cloud computing data storage systems. Cloud storage (eg. Amazon S3) is emerging as a popular service and most of the enterprises shifts their data workloads to the cloud. The popularity of Cloud computing system increased dramatically because it rent computing resources, bill on a pay-as-you-go basis and multiplex many users on same physical infrastructure. Cloud users are provided an illusion of infinite computing resources so that on demand resource consumption rate can be increased or decreased. Cloud computing offers the vision of a virtually infinite pool of computing, storage and networking resources where applications can be deployed. It is the popular solution for on-demand and dynamic resource provisioning. The provisioning & maintenance of cloud resources are done with the help of resource management techniques (RM). The RM techniques are responsible to keep the track of free resources and assign the resources from free pool to incoming tasks. Along with the growing demands of modern applications and workloads, cloud computing has gained prominence throughout the IT industry. Modern applications are growing along various dimensions such as the number of users, complexity, data size and many people are connecting to the internet through various devices. Web Applications which deals with the cloud storage systems contain heterogeneous workload to deploy on the cloud. This heterogeneous workload needs to be handled carefully otherwise problems like long scheduling delay, starvation for low priority tasks may occur which can hurt the performance of the application significantly. To handle this workload characterization is much more important. In workload characterization, heterogeneous workload is divided into multiple task classes with similar characteristics in terms of resource and performance objectives. It is important to consider heterogeneity of workload. The workload typically consists of diverse applications with different priorities and resource requirements. Failure to consider heterogeneous workload, it will lead to long scheduling delay, starvation by affecting performance of the application. Modern applications faces challenges such as workload characterization, resource allocation and security. To meet these needs we propose a new framework which interacts with the heterogeneous database systems and provides workload characterization by using K-means clustering algorithm, resource allocation by providing virtual machines through requesting to the cloud and security by using AES algorithm for encryption and decryption. In this research work, transactions are taken in form of heterogeneous files from users as input and deploy the files efficiently at cloud storage system as output. Transactional operation like insert, update, delete are performed on users data at cloud storage systems. Here, we are attempting to design novel approach for secure, energy efficiency, scalable transaction management system based on terminologies and methods of resource scheduling in heterogeneous distributed database systems. We address “the transaction management problem", which is to interact with the heterogeneous database systems and to allocate and schedule computing resources with workload characterization by providing security to user’s data. Performance Evaluation of the proposed system is done based on fluctuating workloads and in terms of Throughput, Energy Efficiency, CPU Utilization, Scheduling Time, Response Time. The performance of the system is measured and compared with Container Based Scheduling (CBS), Dynamic Power–Saving Resource allocation (DPRA) mechanism and SLA-Aware Energy resource management (SLA) technique. The Experimental results of the proposed system shows effective performance of the system. Results shows that proposed system outperforms existing techniques. The implemented environment is easy to use and the input to the system are given for execution in a normal way without changing or restructuring the code.

Other Details

Paper ID: IJSRDV7I10464
Published in: Volume : 7, Issue : 1
Publication Date: 01/04/2019
Page(s): 840-844

Article Preview

Download Article