Since a smart dust Internet of Things (IoT) system includes a very large number of devices sometimes deployed in hard-access areas, it is very difficult to prevent security attacks and to alleviate bottleneck phenomena. In this paper, we propose a lightweight blockchain scheme that helps device authentication and data security in a secure smart dust IoT environment. To achieve our goals, (1) we propose the structure of the lightweight blockchain and the algorithm of processing the blockchain. In addition, (2) we reorganize the linear block structure of the conventional blockchain into the binary tree structure in such a way that the proposed blockchain is more efficient in a secure smart dust IoT environment. Experiments show that the proposed binary tree-structured lightweight blockchain scheme can greatly reduce the time required for smart dust device authentication, even taking into account the tree transformation overhead. Compared with the conventional linear-structured blockchain scheme, the proposed binary tree-structured lightweight blockchain scheme achieves performance improvement by up to 40% (10% in average) with respect to the authentication time.

No abstract

The Internet of Things (IoT) concept, which involves communication within a network of objects, has become increasingly popular, and development in this area of technology is quite active. IoT environments involve the generation of large amounts of data, and require various levels of Quality of Service. For this reason, message scheduling schemes to deliver the data in IoT environments are considered to be essential.This paper proposes an efficient multi-class message scheduling algorithm for healthcare IoT environments. In the proposed message algorithm, messages are grouped into three message classes (UNC, RT and DT classes) based on the characteristics of the messages. The proposed message algorithm is resourceefficient because it uses a simpler priority calculation than the Multi-class Q-Learning message scheduling algorithm, which we proposed for IoT systems previously.Tests were conducted on a part of the oneM2M-based IoT system we have built. These showed that in most cases, the proposed scheduling algorithm performs better than the Multi-class Q-Learning scheduling algorithm.