ONE Summit 2024

Code is expensive and fragile. Data is plentiful and chaotic. And NFV orchestration involves automating both at scale. A rule of thumb for efficiency and responsiveness is that code should live next to the data it processes. However, the complexity of off-the-shelf large-scale compute and data platforms (“cloud”) has presented the telco industry with architectural challenges for optimization. On the one hand our systems have made hard things easy, while on the other hand they have made easy things hard. Let’s take a look at how Nephio and Kubernetes can help us reduce complexity by enabling simple, straightforward engineering decisions for the placement of code and data. We will look at the choice between management clusters vs. workload clusters, delve into the Nephio specialization pipeline and the Kubernetes operator reconciliation loop, and compare Nephio’s configuration-as-data (CaD) paradigm to the more limited capabilities of Helm charts. With Nephio, we’re essentially breaking the Helm chart apart and putting the pieces where they belong.