Dynamic authentication for intelligent sensor clouds in the Internet of Things

Sensor clouds are formed by IP-enabled wireless sensors and Internet of Things devices that are used for sensing and actuation in commercial and industrial applications. Data collected by the sensors are consolidated by distributed cloud data consolidation (DCS) servers to be utilized as raw sensory information by applications running data analytics and actuation functions. Alternatively, DC servers may feed sensor data to the cloud-hosted Big Data Analytics (BDS) servers. Sensor clouds and their respective DCS servers, as well as BDS servers, may form different security realms. These security realms' ownership structures are complicated and differ from standard database servers, necessitating a dependable authentication technique to provide trusted access to DC and BDS servers. This paper proposes a new multiparty authentication framework to authenticate applications requesting access to the DCS and BDS servers without direct human or application access to the sensors and actuators. Only DC servers are permitted to communicate with sensors/actuators, and only applications certified by a Session Authority Cloud are granted access to DCS/BDS servers via an authentication protocol that includes many information and key exchanges. This solution may assure the reliable deployment of sensor clouds in different critical application domains (i.e., industry, commercial, national security, and defense, etc.) while reducing the potential of direct espionage of sensed/actuated systems. Linear Temporal Logic is used to explicitly analyze and establish the correctness of the presented framework. OPNET modeling and simulations are used to illustrate the protocol's design and operations. The results demonstrate that multiparty authentication is conceivable for Sensor cloud computing systems.