Abstract—Network on Chip (NoC) architectures replace traditional bus-based network topology to support multicore processors to improve communication efficiency. A System on Chip (SoC) integrates several processing elements or the cores on a single chip. NoC is a scalable solution for packet-based networks that enables a higher level of parallelism and also performance. However, the performance can degrade and cause severe system instability due to high network traffic load, mainly due to Denial of Service (DoS) attacks like flooding. To alleviate flooding attacks in NoC designs, this research emphases recognizing and blacklisting malicious nodes responsible for injecting large amounts of packets. There are several ways that the network can be corrupted in NoC. Manufacturing defects can cause faulty boards. If one of the nodes is corrupted by a malicious program, it can inject more packets, flooding the network. The objective of the paper is to detect malicious nodes flooding the network through traffic pattern analysis by injecting an excessive number of packets. The detected malicious node is blacklisted, and all packets emerging from that node are dropped, hence preventing a flooding attack. This technique can recognize nodes instigating irregular traffic loads by examining traffic patterns. The experimental results emphasize that the proposed approach effectively enhances the security and reliability of NoC architectures. It achieves a high packet delivery ratio and maintains a meagerfalse positive rate, demonstrating strong capability in detecting malicious nodes. This significantly improves the throughput while managing network congestion efficiently. This ensures that malicious nodes are isolated more effectively and flooding attacks can be prevented robustly. In general, these results confirm that the proposed approach preserves network performance, thus improving the security of NoC systems.