A Classification Based Approach For Data Confidentiality in Cloud Environment

Abstract

Data security in cloud computing is a hard and tiresome task that has not been completely achieved. Various techniques have been proposed for securing data in cloud. Data encryption is a widely used technique for securing the data in cloud. An accurate data security strategy in distributed computing can be decided by first understanding the security necessities of data followed by the selection of possible approach for securing the data. This will help in deciding which data needs to be secured and which not. This paper presents a data classification technique for data security in cloud environment. An improved bagging and boosting algorithm is employed for classifying the data into sensitive i.e. private and non sensitive i.e. public data. After the data is classified, blowfish algorithm is applied for securing the sensitive data and non sensitive data is sent to cloud without encryption, hence saving the overhead and time for securing the entire data. Moreover for upgrading a secure cloud system, the cloud is divided into segments thus dividing the data and storing it in different segments instead of storing the entire data on a single cloud. Thus this algorithm boosts the security on cloud system. Also the results show that improved bagging and boosting technique gives better results compared to K-NN classification algorithm thus reducing the classification time and enhancing the accuracy.