Pollution attacks are one of the major concerns facing P2P networks. They have a tremendous impact on push-based fully connected overlays, in which each peer receives an exclusive chunk from the source and is also the only one responsible for relaying it to the rest of the peers. In this study, we propose a novel technique to identify and expel malicious peers which involves using trusted peers, software-defined networking (SDN) and proactive moving target defense. Experiments to obtain the accuracy and effectiveness of the implemented methods, as well as an analysis of the performance concerns, were carried out through simulation using a Mininet network emulator. The experiments demonstrate the feasibility of our proposal, which provides high rates of detection, not only in pure SDN environments but also in mixed ones.

污染攻击是P2P网络面临的主要问题之一。它们对基于推送的完全连接的覆盖图产生了巨大影响，其中每个对等方从源头接收排他性的块，并且也是唯一负责将其中继到其他对等方的块。在这项研究中，我们提出了一种新颖的技术来识别和驱逐恶意对等体，其中包括使用可信任对等体，软件定义网络（SDN）和主动移动目标防御。通过使用Mininet网络仿真器进行仿真来进行实验，以获取所实现方法的准确性和有效性，并对性能问题进行分析。实验证明了我们建议的可行性，该建议不仅在纯SDN环境中而且在混合SDN环境中都提供了高检测率。