flywheel energy storage system

The need for smart end-to-end monitoring in 5G

November 22, 2020

5G is changing the way we live, work, and communicate. Technology is already deeply embedded in our lives, and 5G is making this even more so with the introduction of smart cities, grids, billions of connected devices, ultra-fast downloads, and low-latency networks in which everything is connected, all the time.

To deliver these, and many other complex use cases, operators are deploying a new 5G network architecture, which includes a new cloud-native 5G core (5GC) with a new radio access technology (5G NR). This modern mobile architecture is the platform that will deliver these exciting new services and applications. However, with every new technology transition, operators need to gain end-to-end visibility, across their entire network, from the RAN to the Core.

Virtualizing the RAN

Virtualization will aid operators in every sense in their transition to 5G. Virtualization will be more cost-effective and allow operators to manage their services more dynamically, which 5G will require.

The Radio power transmitter is not exempt from virtualization. The digital unit acting as the hardware baseband unit will be virtual, named “vRAN”. Until recently, this was located close to the radio power transmitter on the mast, or below, but with virtualization moves to the network edge and functions on generic servers at a central location. This decoupling of RAN functions from proprietary hardware enables operators to manage their network more efficiently, scaling network resources dynamically, which will be essential for deploying high bandwidth-on-demand 5G services.

This decoupling goes further with Open RAN (O-RAN), as the name suggests, has an open, interoperable architecture, meaning operators are freed from the vendor lock-ins, which it hopes will help speed up 5G network development. Led by the O-RAN Alliance, which includes several major industry players such as Telefonica, AT&T, and China Mobile, have defined two clear principles for the O-RAN. The first is its open interface, which it says will allow some of the smaller vendors to share a piece of the market, and more importantly, will enable operators to pick and choose elements from a variety of vendors making their RAN customizable. This will both allow competition and reduce costs. The second principle is intelligence. The deployment of 5G means significantly increased levels of traffic and data that make it no longer viable for humans to operate. Therefore, as much of the network as possible needs to be open and intelligent for operators to be able to ensure swift root cause analysis and excellent service experience.

The critical role of the RAN

It has been estimated that about 70% of all problems experienced by mobile subscribers are due to issues with the radio. These radio-related issues have a critical impact on the customer experience, including a lack of radio coverage and low signal levels. These affect performance in many ways, but the operators’ main concern will be focused on, dropped calls, and slow data connections. Now, with the introduction of the 5G NR that may or may not be using new vRAN, the access network is a critical part of delivering exciting new services and use cases to customers.

An operator needs the data from the RAN to be able to draw concrete conclusions on the root cause of an issue. Without the RAN data, core metrics such as Packet Jitter, Packet Loss, Timeouts, Throughput, and Connection Release Cause may be collected and measured, but that is where the road comes to a dead end. Operators wanting to measure their 5G customer experience with a user experience score will need to take Core data and combine it with the RAN data in order to gain full network visibility.

More than ever, operators are measured by the Quality of Experience (QoE) they deliver to the end-user. So, the operator must be monitoring the RAN effectively.

A layering effect

5G will create a significant increase in both the quantity of traffic and devices and will create volumes of data that no human will be able to monitor manually. Operators looking to deliver a more comprehensive offering will need to offer an equally complete monitoring system that is able to connect the dots across the entirety of the network, meaning from the RAN to the network core.

To achieve this, operators need to deploy advanced machine learning (ML) capabilities which when combined with an intelligent service assurance solution, creates a layering effect. These advanced techniques use thousands of KPIs and parameters from the network to help find anomalies in the data. Of course, the more information you input into the system, the more refined the anomaly detection becomes, meaning the machine will be guided first by a domain expert for fine-tuning, but then the ML algorithms will learn from its own feedbacks and grow increasingly more astute.

Adding the ML layer alongside a service assurance solution enables the operator to gain insights into KPIs such as Slow connection time, Mobility performance, Radio link failures, and inadequate indoor and outdoor coverage. These are taken from the KPIs on the Air Interface RRC or the X2 N1, N2, and Layer 2 events. ML can then recommend various actions, including:

  • Open a ticket and send out an alarm of the defective element
  • Allow drill levels, with impacted number of users and heat map areas
  • Send out command recommendations to bypass the impact
  • Feed a closed Loop System to help solve the identified impact
  • Proactively alerts to minimize the impact on the end-user
  • Alarm a change of trend over time

Service assurance and end-to-end monitoring

Deploying a containerized service assurance solution with advanced technologies in AI and ML will enable operators to gain full network visibility. It is the service assurance containers monitoring both the RAN and the Core functions, which will smartly collect, process, and correlate data using AI and machine learning. Combined they will deliver the necessary insights to the operator for targeted optimization, drill-downs, and root-cause analysis. Without a fully-cloud native service assurance solution, the operator will not be able to adapt easily to network changes such as dynamic slicing that will create new network core elements. Cloud-native service assurance can easily scale its monitoring capabilities across the network. Only by including RAN monitoring can the operator genuinely understand the root cause of an issue and decide the correct actions required to resolve an issue.

The end goal is always an improved customer experience, and with an end-to-end service assurance solution, which includes AI and ML monitoring in both the RAN and the Core, can operators truly gain full network visibility.

RADCOM Service Assurance is a fully automated, cloud-native and containerized solution for 5G. The combined solution is a dynamic multi-functional solution that unifies both the frontend and backend into a single node. To learn more about RADCOM Service Assurance solution for intelligent, containerized, on-demand, network analysis, and how it delivers full network visibility from the RAN to the core click here.

This blog post may contain forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. To read more about forward-looking statements, please click here.

Share this article
Skip to content