This document summarizes a master's thesis project on improving Internet of Things (IoT) security with Software Defined Networking (SDN). The project involved designing an IoT security architecture using SDN, developing an anomaly detection algorithm, and evaluating the algorithm's performance on a testbed. Key results were that a standard deviation threshold of 10 and window size of 10 seconds detected anomalies with the lowest errors. The architecture detected and mitigated attacks by inserting rules to block malicious traffic. Future work could apply the approach to more sensors and different security architectures.