INTRODUCTION OF AN AUTHENTICATION METHOD FOR SECURING DATA IN HADOOP SYSTEM

Abstract

The increasing popularity of cloud basis Hadoop system has led to the improvisation of security. It permits users for storing and processing purpose of huge data at exceptionally low costs however it lacks security measures to carry out satisfactory authentication as well as authorization of users and services. In our work we introduce a token-based approach which provides more secured protection of distributed file system data without burdening authentication functions. The introduced authentication approach protects sensitive distributed file system data against various attacks such as impersonation and replay attacks. The system will make use of hash chain of authentication keys, rather than public key-basis authentication keys which are found in existed file systems. The scheme allows clients to be verified by data node by the use of block access token while thwarting replay as well as impersonation attacks. The proposed system will include an additional layer of security towards the traditional symmetric key Hadoop’s distributed file system authentication process. The technology of elliptic curve cryptography makes authentication keys unidentified, thus protecting them against various attacks. The scheme allows clients to be verified by data node by the use of block access token while thwarting replay as well as impersonation attacks.