Service replication and flow monitoring on next-generation software defined networks

• Autor(a) principal: Santos, Rodrigo Lopes da Silva
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/38988
• topic: Next-generation Software Defined Networks; Virtualization; Kubernetes; Docker; P4; Monitoring; Controllers

The development of new data plane and control plane technologies for Software Defined Networks, such as the software switches, languages like P4, and control specification APIs like P4Runtime, enable the possibility for very great innovation in this area. Adding to this the ability to have virtualized services in a light but efficient container, brings forth the capability to deploy services easily on the edge of the networks. Also virtualized infrastructure like Kubernetes makes it possible to manage these containers while providing dynamic scaling and management methods. With the ability to reconfigure and update the data plane packet processing, and in data plane computation, the potential for flow monitoring, and service replication request directly on the data plane, rises. In this scope a system was implemented that accomplishes just that, a P4 programmable switch was utilized, in conjunction with a gRPC controller to manage it. Making use of P4 match+action tables, service flow detection and count was developed. The switch is able to do flow counting utilizing data plane registers, and service replication request, through packet cloning and controller communication. Tests were made to study the correctness, and time performance of the system, and its most important individual processes, such as the instantiation times, detection times and deletion times.