Proposed Architecture for Improving Security and Consistency of Data Transactions in Cloud Database using Tree-Based Consistency Approach J.Antony John Prabu, Research Scholar and Assistant Professor, Department of Computer Science St. Joseph’s College, Trichy, Tamilnadu, India- 620 002 Dr.S Britto Ramesh Kumar Assistant Professor, Department of Computer Science St. Joseph’s College, Trichy, Tamilnadu, India- 620 002 Abstract— Cloud computing is one of the important resources in IT industries to provide different services for cloud vendors and clients. Security is the main issue in cloud because data are stored and maintained by third party environment. Cloud computing has lot of issues to maintain transactional data in cloud databases. It needs to maintain ACID guaranties to execute the transactional data. In this paper, Proposed Architecture, Cloud controller, Depth 1 Fixed Tree Consistency (D1FTC) method, Cloud Data Locker and data storage in cloud have been discussed. It also portraits the needs of ACID guaranties, major security levels and ensure consistency of data transactions. Finally this paper affords thorough study on proposed architecture for improving security and consistency of data transactions in cloud database. Keywords--- Cloud DTM, 2-Phase Commit Protocol, Cloud Security, Consistency in cloud DBS, Cloud Controller, D1FTC and Cloud Data Locker. I. INTRODUCTION Cloud computing provides everything as a service and it has different deployment models. It gives reliable service for analytical data but not for transactional data management [32]. Most of the cloud services are deployed in the hybrid cloud environment. The cloud vendors are making a service agreement with cloud providers and launch the reliable services for their clients. Instead of owning the software and hardware vendors move to cloud environment. So the cloud vendors need not worry about the software and hardware maintenances. Cloud has some technical issues to handle the data in distributed environment [10]. The cloud researchers are designing and developing new techniques for scalable transactions to avoid workload consistent[47,49], concurrency controls to improve serialization access [4, 8, 22], consistency approaches for efficient transaction processing system[24,25,26], scalable data storage model for better data handling[7,12,14], data security models[18,19,21] for efficient data handling and security algorithms for multilevel security. Therefore providing security is one of the common and major concerns in cloud, so multilevel security is needed to product each service without loss of data [17, 33, 48]. The next major issue is maintaining strong consistency state in the database level especially for transactional data management services [46]. Hence most of the research works are concentrating to develop the reliable methodology and architectures to strengthen the security for cloud services and strengthen consistency state in cloud distributed database system [15, 16, 30]. II. REVIEW OF LITERATURE R.Anandhi, et all proposed a model to improve the consistency state of the data transaction and also analyze the performance factors of different scalability options in cloud databases. This paper insists that reliability of cloud transaction applications depend on the consistency and scalability levels [13]. Chang Yao, et all proposed a concurrency control protocol named DGCC ((Dependency Graph based Concurrency Control). This protocol is used to achieve better execution of transaction and scalability among database systems. This DGCC based OLTP system also integrated with efficient recovery mechanisms [1]. Pornpan Ampaporn, et all explore a performance analysis against two different data consistency models by leading cloud providers. The results show that writes performance was 3 times worse than reads and it also has grater variance of consistency rate. Hence a better consistency approach is needed for data transactions in cloud environment [3]. Aleksey Burdakov, et all proposed a consistency model for NoSQL databases for data transactions. It explores the characteristics of data consistency and analysis performance of different consistency models [6]. Álvaro García-Recuero, et all proposed a consistency model to efficiently replicate the data among long geographic distance in cloud environment. This approach secures the overloading of both network and system side. The architecture builds with three dimensional vector field models to handle different applications in cloud [5]. Jens Kohler, et all proposed a architecture called data cache architecture with implementation of both parallel and lazy fetch strategies. This work explores the performance analysis between the two strategies and discusses to overcome the issues with SeDiCo framework [7]. Thuy D. Nguyen, et all proposed a prototype called MLS column-store following kernelized design pattern. This approach used in cloud-scale data storage system. It explores the guaranties of efficient cloud-scale data storage in distributed system [17]. Sebastijan Stoja, et all proposed a architecture for realtime database in cloud data transactions. This paper explores the other important topics and analyzes the merits and limitations of it [38]. Marco Serafini, et all proposed International Journal of Computer Science and Information Security (IJCSIS), Vol. 15, No. 12, December 2017 118 https://sites.google.com/site/ijcsis/ ISSN 1947-5500