Mecanismo de prevenção de ataque DDoS em redes SDN

• Autor(a) principal: Vieira, Alfredo Menezes
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFS
• Texto Completo: https://ri.ufs.br/jspui/handle/riufs/15014
• topic: Computação; Rede definida por software (Tecnologia de rede de computador); Proteção de dados; Ataque Distribuído de Negação de Serviço; DDoS; Rede Definida por Software; SDN; Prevenção; Distributed Denial of Service Attack (DDoS); Software Defined Network (SDN); Prevention; CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO

The Software Defined Network (SDN) offers benefits such as scalability, flexibility, monitoring and ease of innovation, due to its main characteristic of separating the data plane from the control plane. Communication between the controller and the data plane is carried out through the OpenFlow protocol, allowing the sending and receiving of messages from a switch that supports this protocol. In this way, it allows the SDN controller to send instructions through codes developed in various programming languages to the network devices. Due to its logically centralized and software-controlled structure, the controller becomes a strategic target in carrying out attacks. Among the many existing threats, the distributed denial of service (DDoS) attack has a destructive effect on SDN networks. The main objective of this cyber attack is for legitimate users to be harmed due to denial of service. The execution of the attack has distinct phases and counts on infected devices which are called bots, forming an army known as botnet. DDoS attack prevention involves methods that aim to prevent the network infrastructure from falling victim to this form of attack. Given the results observed through a systematic mapping, we decided in this work to propose and analyze a mechanism for preventing DDoS attacks in SDN networks that acts in the first phase of the attack, protecting the SDN controller. Of the two types of existing scans (horizontal and vertical), it was observed from the experiments that the engine obtains from 98.64% to 99.37% accuracy, 63.89% to 82.76% accuracy and 77.97% to 84.62% F1-Score for vertical scanning and 99.73% to 100% accuracy, 99.46% to 100% precision and 99.73% to 100% F1-Score for horizontal scanning. It can be useful for SDN network administrators in the context of defending this type of infrastructure.