Why monetisation of 5G Networks requires autonomous operations
With the introduction of 5G networks, the world is becoming hyper connected, albeit virtually. 5G presents a giant leap forward in giving communications service providers (CSPs) the technology to deliver dynamic services with unprecedented levels of automation. The benefits of a full 5G (5G SA) network are immense and include the flexibility to meet the needs of diverse applications and services for latency, bandwidth, security and more across multiple technology and service provider domains, says Ari Banerjee, senior vice president of strategy at Netcracker.
These benefits can translate into new 5G monetisation opportunities, such as autonomous vehicles and automated manufacturing facilities and the efficiency of fleet tracking and stock control to the simplest connected sensors. The capabilities of 5G opens a plethora of potential new revenue streams for CSPs. 5G is the enabler for true augmented reality (AR), virtual reality (VR) and machine learning (ML)-powered services. These in turn are driving the creation of smart cities, smart factories, and smart grids, and catalysing the explosion in demand from consumers for entertainment services such as immersive gaming, and practical enhancement services such as smart homes and remote security.
However, to exploit the paradigm shift that 5G brings to service delivery, the old ways of managing networks will not be sufficient. As well as the leap forward in terms of speed and latency, the unique capability of network slicing means that the 5G network can effectively be multiple logical dedicated networks on a common platform. With network slicing, the 5G network is able to adapt to the external environment, rather than the environment being constrained by the network’s capabilities, as was the case with previous generations of mobile networks. While brilliant, this makes a complex environment even more challenging to manage.
Flexibility is key different use cases have different requirements. Fixed sensors that only transmit are a world away from self-driving cars in terms of making demands on the network, so the architecture for 5G must be capable of monitoring, meeting, and adjusting to all the requirements of its users, whatever they may be, on demand.
The key to managing 5G networks is in optimising resources in order to achieve real-time adaptation based on the different levels of latency and bandwidth requirements across network slices. Achieving this delicate balancing act requires a new microservices-based architecture to support business support systems (BSS), orchestration, and operations support systems (OSS) applications. Complex systems are broken down into loosely coupled architectures with modular, cloud-native components.
Here’s where the benefits of cloud economics come to the fore. Cloud platforms, whether public or private, utilise containers capable of deploying microservices in a matter of seconds for dynamic and scalable adjustment of services and near-instant allocation of resources to meet demand. However, it is essential that OSS applications work in tandem with BSS to ensure the complexities of these heterogeneous 5G networks are properly managed. Cloud platforms enable a partner ecosystem to develop on open architecture with standardised APIs for effective integration of OSS and BSS applications.
Yet this new architecture requires a new management paradigm. To achieve outstanding predictive and reactive network behaviour requires a level of automation that has not been implemented in any other technology generation. It is only through artificial intelligence (AI)-driven OSS and BSS applications that 5G networks will be fully capable of delivering on their promises. In addition, there is also the OPEX issue.
Despite the sophistication of and investment required in new 5G networks, CSP customers still demand top quality service and CSPs still need to meet demanding service level agreements. Trying to maintain those manually is no longer viable from a cost of human resource perspective, and manual processes naturally constrain the speed and effectiveness of the network performance. AI-powered autonomous networks with end-to-end orchestration are a must in order for service providers to monetise the new generation of digital services.
By automating service and network design, provisioning, optimisation and assurance with service orchestration, CSPs can deliver dynamic digital services and network slices – which gives the added advantage of being able to rapidly locate the cause of any issues. This orchestration, particularly across multiple cloud hosts, is a critical step in 5G resource management and service deployment. A further benefit is the flexibility that this gives to CSPs as they negotiate the increasingly complicated landscape of multi-vendor ecosystems.
AI, particularly when combined with advanced analytics and machine learning (ML), provides additional benefits to network and service automation. Utilising actionable, data-driven insights from performance monitoring, fault management, traffic levels and so on, dynamic AI-driven decision making will enable maximum levels of optimisation in a closed-loop environment, without any need for an inflexible rules-based approach to network management.
The goal is intent-based automation, where an implementation or action request such as opening a new smart office hub, or a change request such as updating the global security policy of a company can be made, and orchestration systems are intelligent enough to carry out the requests without human intervention. This ideal world of service management makes operational activity practically a background activity.
At a command, CSPs can adapt to customer needs and execute enhancements rapidly, efficiently and with little if any impact on human resources. Achieving the highest levels of automation is a necessity if CSPs are to take advantage of everything that 5G offers in a cost-effective manner. An ever-widening variety of customers requiring on-demand services, coupled with services potentially hosted on multiple cloud platforms and involving many third parties – all require efficient management and seamless integration. For CSPs, partial automation is not an option.
The author is Ari Banerjee, senior vice president of strategy at Netcracker.