Um mecanismo para coleta, análise e classificação de tráfego em redes definidas por software utilizando estruturas de big data

• Ano de defesa: 2020
• Autor(a) principal: Amorim, Alex de Santana
• Orientador(a): Salgueiro, Ricardo José Paiva de Britto
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Programa de Pós-Graduação: Pós-Graduação em Ciência da Computação
• Palavras-chave em Português: Engenharia de redes
• Palavras-chave em Inglês: Big data; Machine learning; Traffic engineering; SDN; QoS
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://ri.ufs.br/jspui/handle/riufs/14145

A variety of new applications and services appear all the time in the Internet universe, they arrive as a veritable avalanche of information in network infrastructures and need to be understood and treated in their particularities to guarantee the good performance and better use of network resources. Analyzing flows and extracting knowledge to support management decisions and intervene in a timely manner to possible setbacks is a major challenge for network professionals in traditional structures. Given this scenario, the concept of Software Defined Networks (SDN) emerged as a proposal for a more dynamic, manageable and adaptable approach, where data plan and control plan are decoupled, allowing centralized control. This makes the task of obtaining information about network traffic simpler and even makes it possible to use this large volume of data (big data) to extract traffic behavior patterns using machine learning techniques. These standards can be used to characterize and classify traffic and thus apply quality of service policies that can improve the use of network resources. Thus, the main objective of this work is to develop the DETCCS - Decision Engine for Traffic Classification and Control in SDN, a mechanism to assist decision making about traffic in SDN, based on the collection of information from data flows, structures big date, identification of patterns and classes of traffic and application of policies to guarantee QoS (Quality of Service) or improve network utilization. Initially, a systematic review of the literature was carried out in order to understand how big data has been explored to assist the quality of services in SDN. This study allowed to assimilate concepts, identify tools and elements that are commonly used in the literature, as well as the gaps that could be explored for the development of this work. In the case study, the results obtained on the data extracted from a real dataset, point to a segregation in the behavior of flows in 3 groups. They demonstrate a distinction as to the distribution, volume and duration of flows within the analyzed scenario. From the classification of the different flow groups, DETCCS was used to apply rules that could limit traffic to some groups, to the detriment of another that would hypothetically have higher priority. The experiments carried out show a great potential of the proposed mechanism, which can be used by network administrators.