A group key based security model for big data system

Abstract

Recently Big data is in the spotlight and several NoSQL systems have been appeared in order to process large scale data. Among them Cassandra provides high scalability and availability with internal cluster structure. However, it does not provide enough security functionalities, especially transferred messages between internal cluster nodes are easily exposed by outside adversaries. In this paper a group key based security model for Cassandra is proposed. Cluster membership authentication and message confidentiality is provided by using of the group key, and the key is efficiently updated by decentralized approach where the cluster is divided into several subgroups considering Cassandra structure. Our model contributes preventing Cassandra cluster from illegal outside access attempts.