Towards Secure Fog: Establishing A Secure Authentication And Rogue Fog Node Detection In Fog Computing

Fog computing also renowned as Fogging and it is a new and emerging computing technology.It extends traditional cloud computing services such as computing,networking,and storage towards the edge or boundary of the network.It comes up with various features and facilities such as reduce service latency,large-scale Geo-distribution,heterogeneity,real-time and mobility.With the new technology,new diversified challenges arise in the Fog.Security and privacy are significant concerns of this computing architecture due to the distributed ownership of Fog devices in the environment.However,in this dissertation work,we are concentration on various security and privacy issues related to Fog.From the various security issues,we select two different issues to come up with different unique solutions according to the Fog computing perspectives.1.We have proposed taxonomy on the basis of various privacy and security issues related to Fog computing.We have identified various security and privacy issues such as trust management,privacy assurance,authentication,access control,threats and attacks,and security audit.We have discussed and analysis from the various literature and presents their limitation and drawbacks in terms of Fog computing architecture.2.Providing secure authentication between Fog devices at the Fog layer is the key challenge because of the massive scale distributed Fog devices.The traditional authentication methods are not directly applicable due to the unique architecture and characteristics of Fog.Moreover,the traditional authentication methods require more computation power and have high latency which does not meet the key requirements of Fog.To fill this gap,this thesis proposed a secure decentralized location-based device-to-device(D2D)authentication model where Fog devices can be able to mutually authenticate with each other at the Fog layer by using Blockchain.We consider Ethereum Blockchain platform for the Fog device registration,authentication,attestation and data storage.We describe system components,architecture,and design,and we also discuss key aspects related to security analysis,functionality,testing,and implementation of the method.We validated our model by comparing with the existing method and results show that the proposed authentication mechanism is secure and efficient.3.One of the most significant and challenging security concerns in Fog is the presence of malicious or rogue Fog node.Without proper detection mechanisms,it could be difficult to manage the Fog node in a distributed Fog environment.There are renowned techniques for rouge node detection in traditional or cloud computing,nevertheless,this thesis tries to solve the problem from a different perspective.In this thesis,we have introduced a statistical-based approach to detect a rogue Fog node using a Hidden Markov Model(HMM).Our trained HMM can successfully detect the presence of a rogue Fog node instantaneously within very short computation time with high accuracy.We have verified our proposed approach using MATLAB simulated environment.Moreover,the thesis ends up presenting some results assessing the benefits of the presented model on account of performance compared with the existing methods.The results accomplished in the research work shows that the approaching of the authentication technique and rogue Fog node detection technique is secure and efficient.Therefore,the methods are suitable to implement in the Fog environment.Finally,several open issues are identified and discussed in terms of overall Fog security issues as well as authentications and rogue Fog node detection.