Internet of Things (IoT) devices are resource constrained devices in terms of power, memory, bandwidth, and processing. On the other hand, multicast communication is considered more efficient in group oriented applications compared to unicast communication as transmission takes place using fewer resources. That is why many of IoT applications rely on multicast in their transmission. This multicast traffic need to be secured specially for critical applications involving actuators control. Securing multicast traffic by itself is cumbersome as it requires an efficient and scalable Group Key Management (GKM) protocol. In case of IoT, the situation is more difficult because of the dynamic nature of IoT scenarios. This paper introduces a solution based on using context aware security server accompanied with a group of key servers to efficiently distribute group encryption keys to IoT devices in order to secure the multicast sessions. The proposed solution is evaluated relative to the Logical Key Hierarchy (LKH) protocol. The comparison shows that the proposed scheme efficiently reduces the load on the key servers. Moreover, the key storage cost on both members and key servers is reduced.

INTRODUCTION

The Internet of Things (IoT) refers to billions of interconnected smart objects equipped with sensors or actuators. Existing and evolving communication protocols are used to enable services to such smart objects and allow them to communicate among each other and with backend users of the Internet during different activities of sensing and controlling. IoT networks, without intelligence, are just wireless sensor networks. Intelligence in IoT is achieved using context awareness. Context awareness, as defined in [1], is the ability of a system to provide information or services to users using information of a certain entity where the entity can be a person, place, piece of software, or an object. On the other hand, the IoT is naturally a resourceconstrained network regarding CPU power, memory, and energy. Therefore, many of the IoT application scenarios rely on multicast operation to preserve bandwidth and enhance the sensing and control operation among the sensors and the actuators. Multicast transmits data efficiently between one sender or multiple senders to multiple receivers. However, the constrained nature and the massive size of the IoT network make it vulnerable to many security attacks. In order to multicast information among a certain group securely, the traffic should be encrypted. So a common group key should be shared among all members of the group. Whenever membership changes, the group key should be updated. Hence, during the registration process, it is necessary to have strong authentication mechanisms to acquire the identity of the participants prior to distributing the key material. Thus, the main concern is around key management, key distribution, and access control for the key material [2]. Group Key Management (GKM) protocols are divided into three categories: centralized, decentralized, and distributed key management protocols [2][3][4]. Nevertheless, all the conventional GKM protocols under the above mentioned categories don’t suit the dynamic nature of the IoT scenarios and applications. Consequently, many research directions are carried out to adapt these protocols to IoT networks [5][6][7]. Yet, most of the research related to GKM focuses on adapting these protocols by working on just one or two of the IoT aspects such as mobility, scalability, constrained nature of devices, application nature, network access technology, or addressing. They ignore that the majority of IoT scenarios need to work on most of these aspects combined together and they lack the ability to address the resulting issues of such combination.