An efficient approach for device identification and traffic classification in IoT ecosystems

摘要: Internet of Things arises as a computational paradigm that promotes the interconnection objects to and enables interaction, operational efficiency, communication. With increasing inclusion in network intelligent have characteristics such diversity, heterogeneity, mobility low power, it is fundamental develop mechanisms allow management control. In addition, important identify whether assets are working properly or anomalies. Traffic classification techniques aid analysis handle many other key aspects security, management, access control, provisioning, resource allocation. order promote identification devices, especially IoT, this article presents technique uses Random Forest, supervised automatic learning algorithm, together with inspection contents packages for purpose. Also, we use same algorithm perform traffic. end, devices showed an accuracy approximately 99%.