Part 2 of our series from our joint whitepaper with Analysys Mason discusses the impact 5G will have on monetization systems.

The lack of clarity regarding which 5G use cases will be successful will have an impact on CSPs’ ability to formulate investment strategies for monetization systems. Given the large number and variety of these use cases, CSPs will need to re-focus from an approach centered on use-cases to one focused on architectural agility for their charging systems. CSPs transforming their monetization systems to thrive in a 5G world will need to adopt a microservices-based architecture, and be prepared to embrace a high degree of automation.

Adoption of microservices-based architecture

Traditionally, software applications have been designed as sets of tightly-coupled functions that necessarily function as monolithic entities. Such applications are unable to make the most of the flexibility and resource utilization potential of the cloud. This is because their architecture does not enable the differentiated and highly-efficient scaling of individual components, or the agile scheduling of components to maximize performance and minimize consumption. By contrast, applications in a cloud-native architecture are developed as microservices that are mostly stateless and loosely coupled. Larger systems can be composed from microservices, making use of their ability to operate in different contexts. Key characteristics of cloud-native computing include horizontal scalability; reusable, loosely-coupled and open components; continuous integration and delivery; and container-based deployment. These characteristics result in the highest levels of efficiency, resilience and speed when delivering software in the cloud.

Key advantages of microservices architecture over monolithic frameworks

Microservice-based architecture has significant implications for monetization systems such as supporting extensive automation; allowing individual functions to scale independently on demand; eliminating single points of failure and allowing for continuous delivery and integration. These capabilities not only address some of the most pertinent drawbacks of legacy monetization systems, they also enable CSPs to support 5G-specific requirements related to ultra-low-latency, high-throughput and distributed architecture.

Strong emphasis on automation

With profit margins declining and costs rising (Figure 4), CSPs are looking to AI and automation to accelerate service development and execution of operational processes, allowing them to respond more quickly to changes in an increasingly competitive market. As processes become more complex and the associated data sets get larger, it becomes increasingly difficult to derive insights without assistance from AI. Automations based on simple rules are no longer sufficient and too slow in development, particularly in a fast-moving market with demanding customers. A high degree of precision is needed to fully automate human interactions.

AI is central to this transformation because it can be deployed to interpret real-world data to provide efficient and precise interaction with humans, as well as faster development cycles. It also can support increasingly complex processes (beyond applications-based automation) to optimize end-to-end business processes.

Role of AI and automation in improving efficiency

Services-based architecture will reshape telco charging and monetization functions

5G has the potential to transform the telecommunications landscape, but in order for CSPs to be able to offer the range of new services made possible, the underlying network and software infrastructure must change. Legacy point-to-point architectures rely on multiple non-standard and often proprietary interfaces to connect different functions. This makes the architecture highly complex in the long term and limits its ability to scale and share its services freely. Furthermore, the legacy framework typically has dependencies and constraints that make it incompatible with cloud infrastructure.

To realize the potential of 5G, CSPs must implement an SBA. 3GPP defines SBA as a set of interconnected network functions that are preauthorized to gain access to each other’s services. SBA enables the functions to become more granular and independent, improving operational efficiency. It is also vendor- and technology-agnostic, which improves modularity and reduces time to market while also making the architecture future-proof.

The adoption of SBA for 5G will have a profound impact on monetization systems as well, including the charging system. As part of the SBA in 5G, the legacy offline and online charging systems have been combined to form a single CCS. The interface with the billing system has also been updated with new information related to 5G charging.

The CCS is at the heart of the new architecture’s monetization capabilities and includes four functions:

• Charging function (CHF) is the primary function of the CCS. It has been introduced by the 3GPP in order to charge for network events. CHF collects information on all network events over the Nchf interface, which is a services-based interface exposed by the CCS. This is quite a significant shift from data flows for previous generations, where event records were generated by the network switches with detailed information on all events. The CHF can include both the online charging function (OCF) that is used for online charging and charging data function (CDF) that is used for offline charging functions.
• Rating function (RF) is responsible for applying the relevant rates to events processed from the CHF. The RF can be a stateless function as it is agnostic to the device, user and session.
• Account and balance management function (ABMF) keeps track of the users’ credit and debit balances. Importantly, it provides support for both monetary and non-monetary balances and allows multiple balances to be associated with a single account.
• Charging gateway function (CGF) is responsible for aggregating event data records and pushing them into the billing system in a pre-agreed format. The generation of event data records is based on a number of parameters which can be defined separately.

Download the full whitepaper here.