5G SA Network Visibility Challenges and Current Alternatives
According to the GSMA report ‘2022 The Mobile Economy’, there were 8.3 billion SIM connections (excluding licensed IoT) in 2021, expected to reach 8.8 billion connections by 2025. While 5G made up only 8% of those connections in 2022, it is forecasted to jump to 25% by 2025. Initially, Communications Service Providers (CSPs) rolled out 5G non-standalone (NSA), which relies on existing 4G infrastructure. However, 5G standalone (SA) is gaining traction, with 22 commercial deployments by the end of 2021 and an estimated 66 live 5G SA networks by 2023, according to STL Partners.
What is 5G SA?
5G SA represents the next generation of mobile networks that promises to unlock the full potential of enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC), and massive IoT use cases. Unlike 5G NSA, which uses a hybrid of 4G and 5G, 5G SA introduces an entirely new core architecture defined by 3GPP. This architecture is based on Service-Based Architecture (SBA) and adopts a cloud-native software approach, enabling dynamic, scalable, and flexible deployments.
The key advantage of 5G SA is its ability to offer greater network efficiency and improved user experiences by harnessing these features. However, this evolution brings unique challenges, especially in terms of network visibility.
5G SA Network Visibility Challenges
In previous generations, such as 4G and 5G NSA, monitoring network traffic at the packet level was relatively straightforward. CSPs could capture and correlate control-plane and user-plane traffic, including subscriber identity (often hashed for privacy), to generate valuable insights for Customer Experience Management (CEM) systems. These insights helped CSPs understand hotspots, service usage, bandwidth consumption, and user behaviour.
However, due to its complex and distributed cloud-native architecture, 5G SA introduces significant visibility challenges. While open-source tools like Prometheus (for metrics), Grafana (for logs), and Jaeger (for tracing) can provide some insights, they fall short in offering subscriber-level visibility—such as which services a subscriber is using at a particular time or location.
Subscriber Monitoring Solutions for 5G SA
In 5G NSA, CSPs could physically tap a link to extract packets for analysis. However, 5G SA complicates data extraction, as Core Network Functions (CNFs) communicate using encrypted HTTP2 messages. Additionally, 3GPP has not standardized a method for mirroring CNF messages in the same way it did for 4G networks.
The Network Data Analytics Function (NWDAF) was introduced to address this challenge. NWDAF is designed to streamline core network data analytics, generating actionable insights and enhancing the end-user experience. However, its implementation is optional, and not every CSP will deploy it in their 5G SA network.
NWDAF faces several challenges:
- Data availability and granularity for subscriber monitoring may be insufficient.
- It may impact network performance, as CNFs must continuously report to NWDAF.
- Integration complexities and alignment with 3GPP standards pose additional obstacles.
Despite these challenges, NWDAF could offer solutions for network analytics, but its effectiveness in providing subscriber behaviour data remains uncertain.
Current Alternatives for 5G SA Visibility
Some CSPs are exploring alternatives like Service Communication Proxy (SCP) to overcome visibility challenges. SCP acts as an intermediary between CNFs, often deployed as part of a service mesh (e.g., Envoy), providing the ability to mirror traffic. However, this approach introduces additional latency and resource demands, making it less than ideal for real-time monitoring.
Other network vendors propose copying the message at the CNF level before encryption. This data is then streamed to a data extractor, which forwards it for analysis. These implementations vary significantly across vendors, with some using TCP, GRE, or even HTTP2 and JSON-based payloads for transport.
A more cloud-native solution involves using mirrored messages in decrypted form from CNFs, sometimes employing eBPF technology to capture data before encryption. However, these methods also raise concerns regarding data security outside the 5G SBA.
Conclusion
The shift to 5G SA represents a significant technological leap, bringing enhanced security features and a cloud-native architecture. However, monitoring subscriber behaviour and ensuring optimal service performance remain critical challenges. While NWDAF and proxy-based solutions offer some answers, the lack of standardized solutions for extracting decrypted messages hinders widespread adoption.
As 5G SA gains more traction, the demand for more streamlined, standardized approaches to data extraction will grow. CSPs and vendors hope for a solution that offers the necessary granularity for monitoring subscriber behaviour without compromising security compliance with GDPR and other regulations.
For now, CSPs must navigate a complex landscape of multi-vendor CNFs, diverse data extraction methods, and the challenges of encrypted communication in a 5G SA environment.