Publicação
Latency sensitive services in edge clouds
| Resumo: | Business environments that rely on Voice over IP (VoIP) for Real-Time Communications (RTC) has traditional deployments that are suboptimal for the future. These traditional networks struggle to meet latency and jitter requirements. An architecture were designed to address the requirements of redundancy, orchestration, and observability in cloud environments and following Software-defined networking (SDN) and Network Functions Virtualisation (NFV) network architectures. This architecture has a control plane, multiple instances of the telephony software, and a metrics probe. We implemented that architecture with a central control plane, two instances in different cloud providers where Open Source MANO (OSM) instantiates a Kubernetes cluster where we deploy the telephony software using Kamailio and FreeSWITCH as Session Initiation Protocol (SIP) server and media server, respectively. We also deploy the metrics probe, to captures and calculates the call metrics in a sidecar of the FreeSWITCH Kubernetes’ pod. The metrics probe sends jitter, packet loss and Mean Opinion Score (MOS) metrics to our Prometheus server which makes them available to other uses and visualisation in Grafana. We successfully manually tested our system, making calls. Then we automatically tested the system with multiple calls being created and verified the overall performance of the network and the metrics with the multiple concurrent calls. With these tests we validate the redundant, and orchestratable system by observing the metrics. |
|---|---|
| Autores principais: | Alves, André Miguel Sampaio |
| Assunto: | Cloud Edge Kubernetes VoIP RTP SIP Latency Jitter MOS |
| Ano: | 2024 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Aveiro |
| Idioma: | inglês |
| Origem: | RIA - Repositório Institucional da Universidade de Aveiro |
| Resumo: | Business environments that rely on Voice over IP (VoIP) for Real-Time Communications (RTC) has traditional deployments that are suboptimal for the future. These traditional networks struggle to meet latency and jitter requirements. An architecture were designed to address the requirements of redundancy, orchestration, and observability in cloud environments and following Software-defined networking (SDN) and Network Functions Virtualisation (NFV) network architectures. This architecture has a control plane, multiple instances of the telephony software, and a metrics probe. We implemented that architecture with a central control plane, two instances in different cloud providers where Open Source MANO (OSM) instantiates a Kubernetes cluster where we deploy the telephony software using Kamailio and FreeSWITCH as Session Initiation Protocol (SIP) server and media server, respectively. We also deploy the metrics probe, to captures and calculates the call metrics in a sidecar of the FreeSWITCH Kubernetes’ pod. The metrics probe sends jitter, packet loss and Mean Opinion Score (MOS) metrics to our Prometheus server which makes them available to other uses and visualisation in Grafana. We successfully manually tested our system, making calls. Then we automatically tested the system with multiple calls being created and verified the overall performance of the network and the metrics with the multiple concurrent calls. With these tests we validate the redundant, and orchestratable system by observing the metrics. |
|---|
Atividades financiadas
Não foram encontrados projetos financiados.