Maximizing the 5G analytics dividend

Telecommunications service providers have been heavy users of analytics for decades. From calculating subscriber churn to micro-targeting offers, network planning to quality-of-service optimization – data is core to operations. However, the roll-out of 5G networks changes the stakes for the use of analytics by telcos. Not only is 5G the first ‘generation’ to embed analytics as a core function of the network architecture, but it is a vital component in delivering and effectively monetizing the new cloud and software-defined services that are the true promise of the technology. But simply having analytics as part of the standard is only the first step; telcos need to adapt their data architectures to leverage the granularity, scope and scale of data to deliver value. Business leaders as well as technicians need to understand how to use these data assets to maximize returns.

NWDAF – the 1st step

The Network Data and Analytics Function (NWDAF) is a fundamental element of 5G and crucial to delivering a higher level of intelligence to the network. Defined by 3GPPP and ETSI among others, the NWDAF standards are essential to the way 5G operates. For example, the network slicing which underpins many of the use-cases currently being explored and rolled out to individual clients, depends on complex analytics within the NWDAF. Network data analytics are required to monitor service quality and provide failure predictions to pro-actively prevent network and service degradations impacting customer experience. Every individual network slice will have to be monitored for performance and to allow resources to dynamically adjust to meet the specific SLAs for each customer. The complexity of these tasks and the real-time data at scale required for the AI to accomplish them goes far beyond the capabilities of existing BSS and OSS within most telcos.

Cloudification advantages

5G is also a cloud accelerator, and a crucial component in the ongoing digitalization of virtually every aspect of the economy and society. For telcos to play a role, they must embrace the virtual, software-defined, as-a-service models of the cloud. To manage the predicted explosion in connected devices – tens of billions in the next few years – and the associated surge in demand for capacity, networks must transform to be able to deliver faster speeds, lower latency, and more capacity. At the same time increased complexity, range and type of applications, multi-vendor, multi-network environments all threaten to raise operational costs to unsustainable levels. AI and machine-learning algorithms will be essential to automate and manage as much of the network as possible. NWDAF is an essential step in bringing disparate data from across the network together, but it needs to be supported with hyper-scalable data analytics platforms capable of matching the scale, variety and speed of this data.

Automate concept-to-cash

To maximise return on the very significant investments in 5G infrastructure, telcos will want to quickly develop, roll-out and monetise new services that make effective use of 5G capabilities. Beyond higher-speed downloads there are huge opportunities for innovative new services to tap into consumer and business user requirements. A subsequent blog will look specifically at the need for experimentation in 5G, but the capacity to closely analyze network data to spot new behaviors, and uncover new un-met needs is a third important use of data embedded within the 5G network. Using data to not only spot opportunities, but as the fuel for automated ‘concept-to-cash’ processes that create, deploy and monetise services on an as-needed basis will drive telco profitability.

Unify, orchestrate, analyze at scale

5G analytics plays a fundamental role in all these cases; service delivery, cloudification and new service innovation and deployment. But to deliver the insights, and successfully implement new analytic models across telcos so that the business units, operations and support teams can access and action them, requires robust, scalable and shared data platforms. These platforms must be able to unify, orchestrate and analyze network data from the NWDAF. They should also combine data from across vertical applications including BSS and OAM layers within the business as well as external data as diverse as meteorological, transport and local events/news for example. Only then will they really support full automation of closed loop assurance and management of end-to-end services that not only match the promise of 5G but do so in a cost-effective ad commercially viable manner.