With the tremendous increase in the number of Internet of Things (IoT) devices connected to networks, attacks on vulnerable IoT targets are also on the rise. Lightweight devices like IoT have limited battery lifetimes and memory capacities; thus, efficient resource utilization is essential. Existing security solutions require significant memory space and energy to process complete log data, which can lead to network service availability issues. Therefore, a new lightweight security solution is needed for such environments. This study presents a lightweight detection method for random permutation attacks in IoT environments. This method achieves detection by employing long detection intervals when a system is in a normal state. When there is suspicion of a potential attack, it adjusts the detection interval to be shorter, thereby conserving energy and reducing memory requirements for operation in the normal state. We experimentally verified the proposed approach; the results revealed that, compared to conventional methods, it exhibited almost no degradation in detection performance for brute-force attacks, while significantly decreasing the amount of data and detection time.