The social Internet of Things (Social IoT) introduces novel ways to enhance IoT networks and service discovery through social contexts. However, trust-related attacks raise significant challenges with regard to the performance and reliability of these networks. In fact, malicious users exploit vulnerabilities to spread harmful services, necessitating the incorporation of a trust management mechanism in the Social IoT network. To tackle this challenge, we set forward an innovative trust management mechanism empowered by blockchain technology. Through this integration, we aim to mitigate trust-related attacks and establish a secure environment for end users. Additionally, we introduce a new consensus protocol called real-time trust-related attack mitigation Protocol using Apache Spark (ProtoSpark). This protocol which leverages Apache Spark’s distributed stream processing engine to process real-time stream transactions. In the implementation of ProtoSpark, we have developed a new classifier utilizing Spark Libraries. This classifier accurately categorizes transactions as malicious or secure, enabling the protocol to make informed decisions regarding transaction validation. Our research corroborates the superiority of our classifier in terms of predicting malicious transactions, surpassing previous works and other approaches in the literature. Furthermore, our new protocol exhibits improved transaction processing times, enhancing the efficiency of the network.