P2P computing for trusted networking of personalized IoT services

As the development of Internet of Things (IoT) technology has enabled various forms of intelligent services to be provided personalized intelligent services are provided for each person based on the data collected from IoT devices through P2P networking. Intelligent IoT services are gradually expanding. However, there can be various security risks in the devices that make up IoT networking. Untrusted devices can affect personalized IoT services by distorting personal information in analyzing the collected data. Therefore, services by the untrusted devices should be restricted. In this paper, the reliability is defined as the familiarity score, which is determined by the connection experience of the devices in a P2P IoT network. The IoT network composed of devices with high familiarity scores can be defined as a trusted area. In the trusted area, data generated by all devices is used to create knowledge for personalized intelligent services for users. In contrast, personalized intelligent services in untrusted area are restricted. Data generated by untrusted devices is classified by a learning algorithm such as logistic regression; thus, bad data is blocked by a gateway to avoid information distortion in data analysis for personalized intelligent services. The proposed method provides a trusted networking environment to IoT service users and protects data integrity. Thus, it can improve the user's quality of experience (QoE) for personalized intelligent services. The proposed approach is evaluated by computer simulation and its superiority is validated.