Issue link: https://www.e-digitaleditions.com/i/1476549
October/November/December 2022 I 19 Analytics QoE QoS Visibility Gap Observability metrics, traces and logs Visibility KPIs, KQIs, xDRs, traces Services Network K8s, OpenShift, … Infrastructure SoC NoC CloudOps ITOps Observability Visibility MIND THE GAP One of the key benefits of a cloud-native network solu- tion is the abstraction of the hardware infrastructure that underpins the virtualized network and service layers hosted in the cloud. This abstraction relieves developers of the tedious and often time-consuming challenge of adapting their software solutions to various operating environments, both hardware and operating systems. It also enables these networks to easily scale horizontally to address unexpected spikes in demand or uptake of new services. Additionally, the software-based nature allows the development of new and innovative customer services in days or weeks, rather than months or years. Further- more, the service-based architecture and containerization of microservices promote a more open solution where best-in-class network functions can be combined through open application programming interfaces (APIs) to deliver a superior customer experience, while at the same time preventing vertically integrated vendor lock-in solutions. The challenge of this network solution is managing the complexity and scale expected, spurring the need for automation and orchestration driven by ML and AI, the so-called data-driven network. Each NF may be comprised of several 10s of containerized network functions (CNFs) and each CNF may be made up of even more micro- services. Managing even the simplest NF may require monitoring hundreds of performance indicators. For the network and service cloud layers, this data comes from key performance indicators (KPIs), a service assur- ance solution. For the hardware cloud infrastructure, this data comes from an observability solution (e.g., telemetry, traces, logs). The problem is that service assurance solutions are typically the domain of network operations (NetOps) and service operations (SrvcOps) teams, while the observ- ability solutions are typically the domain of the cloud operations (CloudOps) and IT operations (ITOps) teams. FIGURE 2: The visibility gap. To make matters worse, they use different tools and metrics to manage their respective domains, making cooperative analysis of a customer's QoE issues chal- lenging. The hardware abstraction along with the highly dynamic nature of the network and service layers makes it extremely difficult to connect QoE issues in cloud-native services to specific infrastructure hard- ware issues, creating a visibility gap (Figure 2). This gap can lead to extended MTTR times due to incomplete information, miscommunicated support needs, and in some cases, finger-pointing instead of problem solving.