Security Enhancemnt Based On Hybrid Encryption Approach For Cloud Computing

Abstract

— With the advancements in the technology, the amount of data over the internet is getting increased and it
becomes tedious task to handle such a huge amount of data. It is noticed that the clients need to hire the data mining
tasks from cloud service providers this is known as outsourcing. Mostly the selection of cloud service provider for
various services relies upon the choice of clients because the maintenance of confidentiality and privacy is major
concern nowadays. This study specifically conducted to develop to a novel approach (DSEkNN i.e. Data Split Encrypted
kNN) for managing the security by using outsourced kNN (k-Nearest Neighbor) by distributing dataset randomly to
different CSPs on the basis of their limits for handling data. For the purpose of security two-level or multi level
encryption is applied by using RSA (Rivest-Shammir Adleman) and DNA. In DSEkNN firstly the data is encrypted with
RSA and the encrypted data is further crypted with DNA hence it provides two level security to the datasets. For proving
the efficiency of the DSEkNN is compared with OCkNN and PPkNN. After observing the results it is obtained that the
proposed work is capable to decrease the burden of data owner’s by reducing the computation time and communication
cost. The simulation is done on two datasets i.e. red wine and white wine.