To provide the agility needed to support innovative and differentiated offers, the core network is migrating from proprietary physical network elements to hybrid networks with virtual network functions (VNFs) to cloud-native functions (CNFs), containers, and microservices. The end goal is to virtualize everything, including creating virtual versions of physical devices.

Standards plus open architectures and APIs mean that the market is accessible to a wide variety of vendors instead of a few major players, so interoperability challenges will rise to a new level.

5G opens the door to many new capabilities and revenue streams, but the complexity of its networks and operations are raising the industry bar for ensuring reliability, performance, and security.

Of course, for the foreseeable future there will be both 4G and 5G networks, as well as older vintages. That means a hybrid mix of physical, logical, and virtual entities need to be managed and assured, across physical and cloud network domains, from edge to RAN to core.

How service assurance must change

Traditional service assurance approaches must be completely rethought to enable the business opportunities created by 5G technologies.

Fundamental changes are needed to marry the customer-facing business with network-facing operations to identify proactively how network issues impact the service and the customer. Operators must predict, prevent, and act quickly to remediate service-impacting issues that cause downtime or performance degradation—which ultimately impact the customer experience.

5G assurance must be unified and holistic. The ability to ingest data from multiple domains—ignoring nothing—and to normalize the data into a common format can bring visibility to areas that may previously have been lost because of disconnected siloes. Unification and normalization of the different data sources enable the data to be processed holistically based on a single source of truth. This is critical for achieving greatest insights. Machine learning (ML) can be applied for advanced correlation, root cause analysis, and automation across domains—in real-time and at scale.

With 5G, service assurance needs to provide end-to-end visibility across physical, logical, and virtual network entities, across all network domains and vendors. There must also be vertical visibility from the network infrastructure all the way up the stack layers to the services. Both perspectives must be analyzed together automatically to provide accurate insights into network issues and their correlation to services and customers.

The 5G assurance scope goes well beyond 4G and includes traditional physical networks, virtual RAN nodes, transport network interfaces, containers-as-a-service, controllers, network functions such as AMFs and CNFs, resource management VIMs, Kubernetes clusters, and many more.

Topology views must be updated in real-time—becoming a real-time inventory. This live view of the end-to-end and top-down network must be presented on a single pane of glass. The views should be enabled by multiple data overlays and visualized as a map, a hierarchy, or similar, with drilldowns as desired. Issues should be prioritized based on customer impact and indicated in the topology views.

This task is made even more complicated with 5G because of the large number of vendors involved, each with their own agile schedules for frequent solution updates that are driven by continuous integration/continuous delivery (CI/CD) software processes. The network and operations are continually changing in 5G.

As well, 5G brings an exponential increase in the number of devices and the way they are connected within increasingly complex environments. Operators know that these changes are only going to accelerate and become harder to manage.

The need for proactive, real-time service assurance that is driven by actionable insights and event “noise-reduction-through-correlation” requires automation of fault identification, performance monitoring, and root cause analysis. Closed-loop automated assurance is ultimately what will make it all work at scale.

Once an issue’s root cause has been determined, automated workflows should be triggered for quick remediation. Open APIs are essential to enable seamless integration with orchestrators to close the loop by taking remediation action in the network—for example, remediations like increasing container capacity or off-loading traffic from a failing element.