By: Ross Cassan

It’s almost become a telecom tradition for operators to push a great new offering into the market, only to eventually struggle with maintaining quality and delivering a great customer experience as the service matures. The stumbling blocks were always going to be there, but because service assurance wasn’t planned from the start, operators kept themselves in the dark. By the time they were able to turn on the light with a proper service assurance strategy, upwards of a year or more may have passed. And historically, operators have paid the price in frustrated consumers.

Things will be different with network slicing on 5G and all the mission-critical services it will be expected to support. The stakes will be much higher, necessitating that service assurance be baked-in from the start.

Operators are planning now how 5G networks will be sliced to meet customized performance needs in an expanding range of verticals. While approaches will vary by region, stringent demand for service levels unlike anything they’ve had to provide before will be a constant. Missing the mark could mean regulatory fines and SLA violations that eat significantly into revenues. Even more serious, however, is the potential fallout from issues that affect emergency services, public safety, smart cities or autonomous cars, just to name a few.

Guaranteeing great experiences in networks that constantly change

Spirent has been working closely with device makers, network equipment manufacturers and operators on early 5G testing. Working across the full ecosystem has given us a unique vantage point for triangulating expected issues. We see immediate areas of concern at this stage:

• The dynamic nature of network slicing insists the network be continuously reconfigured on the fly to meet the ultra-low latency, high-bandwidth demands of specific applications. Every time this happens, the network is being changed in significant ways. It’s a process that will happen over and over in 5G networks.
• The core network, which used to sit in stable, centralized data centers, now may be pushed to the network edge to deliver enhanced performance. Complicating matters, the control plane and user plane are now separate and independently configured. The distribution of network functions and control user plane separation (CUPS) are radical changes, introducing a new layer of complexity that inevitably results in performance issues and configuration challenges.

Previously, turning up new network infrastructure was a manual practice that could take weeks or months. Lab-based validation was often the long pole in the tent, requiring the use of specialized physical test probes and manual processes. In 5G networks, testing will need to be dynamically triggered, occurring in seconds or minutes. That means it must become software-based, take place in the operational network and be part of an automated workflow. In other words, the testing conducted to assure services needs to be part of how the service is designed.

We know that some infrastructure providers will shout from the rooftops that they plan to handle assurance and QoE requirements in the network gear they deliver. But the simple fact is if these built-in telemetry capabilities aren’t testing the end-to-end network slice performance—right before it goes live for the customer—providers can’t be sure critical SLAs will be met. The industry has not quite woken up to this fact yet and our cutting-edge work in the lab with 5G tells us there will absolutely be issues, from missed latency targets and lack of resources, to misconfigurations and beyond. These potential issues can only be solved with automated, proactive testing that can detect fleeting flaws as they happen, isolate responsible network functions and provide actionable guidance to network management and orchestration layers.