Monitoramento de memória para aumento da segurança em redes: uma proposta em redes definidas por software

Abstract

This thesis proposes a new methodology for memory monitoring and analysis in network devices, with an emphasis on Software-Defined Networks (SDN). The methodology combines memory extraction modules and novelty detection to identify unauthorized changes, such as the insertion, modification, or removal of flow tables. The proposed solution has demonstrated the ability to detect suspicious modifications, isolating compromised devices and redirecting traffic through secure routes, ensuring network integrity. Additionally, the methodology offers an extra layer of security that complements traditional solutions. The approach was tested in simulated scenarios, proving to be scalable and adaptable to different SDN environments. The results indicate that the solution can increase the resilience and reliability of the network by integrating forensic memory analysis directly into the network devices. One of the contributions of the proposed implementation model is that it enables the automation of incident response, enhancing real-time reaction capabilities against attacks.