Abstract
Kubernetes (K8s) has emerged as the most relevant open-source containers orchestration technology enabling
automated computer application deployment and management. K8s is built on the concept of Pod (i.e., a unit of
deployment) consisting of one or more containers that are co-located on a same host. The Kubernetes platform is
based on a flat network structure that enables Pods to communicate with each other on their hosting K8s cluster.
Such flat K8s network, also called Pod network, does not account for network constraints in terms of limited
bandwidth or bounded latency. For this reason, deploying K8s in edge computing environments over metro optical
infrastructures to serve latency/QoS-critical applications requires a specifically designed and
comprehensive framework. In particular, specific workflows are needed to efficiently interface K8s with various
components such as Software Defined Networking (SDN) controller, Service Level Agreement (SLA) broker, and
Telemetry Collector. So far, only few preliminary studies have addressed the K8s integration with SDN control 4-6,
while the overall framework also encompassing the SLA broker and the Telemetry Collector is yet undiscussed.
In this work, we present a comprehensive framework enabling the K8s scheduler to interact with the SDN controller
for deploying Pods and services in metro optical scenarios, taking into account network constraints. The framework
also includes a telemetry system enabling effective SLA monitoring and enforcement. Finally, the framework
encompasses an SLA broker interfaced with the telemetry system triggering the SDN controller to perform automated
network adaptation upon detection of network performance degradation.
Users
Please
log in to take part in the discussion (add own reviews or comments).