Security Analysis of SDN Applications for Big Data

Big Data is a term that describes structured and unstructured large data sets. One of the frameworks to store , process and analyze this data is Apache Hadoop. Software Defined Networking (SDN) enhances the performance aspects of Hadoop by optimizing bandwidth utilization and improving network management. Security attacks on the SDN controller and switches can compromise the whole Hadoop system, that may cause loss or manipulation of valuable data. We selected the three most advanced approaches that focus on accelerating the data transfer between the cluster nodes. FlowComb , Pythia and Hadoop-Acceleration (Hadoop-A ) focus mainly on optimizing performance but do not consider any security aspect in their design. This motivates us to analyze the security aspects of these SDN applications. This paper focuses on the analysis of security features with STRIDE threat modeling technique. All approaches need improvements to gain security. We find that Pythia is natively the most secure approach while other approaches can be secured by deploying add-on security mechanisms.