A zero-touch operational model is becoming increasingly critical for communications service providers (CSPs) to realize the monetary benefits of multi-vendor digital services such as dynamic 5G and Internet of Things (IoT) offerings. However, considerable transformation is required for the CSPs to evolve the service layer of traditional operational support systems (OSS), which is where service management and orchestration is vital.
Rapidly evolving market needs are impacting virtually all aspects of CSPs’ IT and network infrastructure. We are already witnessing the exponential growth of these networks in both size and diversity, making end-to-end (E2E) service orchestration a necessity. The need to deploy new features and services quickly and seamlessly, roll out upgrades, and scale capacity on demand will only increase as 5G networks become more widespread. In addition, 5G’s unique ability to offer network slicing means CSPs will need to provide different levels of service, while meeting strict SLAs.
As network functions evolve to become cloud-native, additional complexity results with the melding of the virtual with the physical elements, further complicated by the increasing number of vendors involved in the network. On top of this, IT systems are also being transformed, so a mediation layer is needed to provide a bridge between legacy and new business support systems (BSS) and operation support systems (OSS) environments and core, edge and transport network domains.
To release the full potential of 5G networks and multi-cloud platforms however, service orchestration needs to provide a real-time E2E view across all technology and cloud domains with a single pane of glass management. With the right level of automation, CSPs will have the tools to build successful digital businesses on top of the new 5G networks. Automating all aspects of cross-domain services, however, requires some key design principles to ensure multi-vendor collaboration.
A powerful approach is intent-driven orchestration – the automatic completion of business requests utilizing dynamic service models, policy and context, with closed loop control. This means that the target state is achieved without implementation details which in turn automates the full service and network slice lifecycle across hybrid networks.
Service automation needs to be complemented with Artificial Intelligence (AI) and advanced analytics to ensure not only effective service assurance but efficient anomaly detection, forecasting and optimizations. The ability for the network to self-learn, reason and adapt control loops means actual issues are detected quickly and rectified; potential issues predicted and addressed; the network is optimized at all times, and changes implemented without the need for human intervention.
Open APIs and standards are also critical to ensuring that the goal of an open and multi-vendor environment does not give way to proprietary systems and vendor lock-in that stifle innovation. This means that the service orchestration must support open APIs and standard service models. This will simplify the integration and automation of multi-vendor services and enable business and network domains to have access to relevant information.
As an example, in the transport network there will typically be several vendors for IP/MPLS and optical networks – each with multiple layers and transport domains. Domain service orchestration will have a complete E2E view of the domain, enabling it to automate service and slice provisioning across this complex mixed technology and vendor landscape. However, domain service orchestration has no visibility or knowledge of services in other domains, or services that cross domains. Real-time information is shared between domains using open APIs.
Implementing these key design principles will then open the door to 5G edge applications and Industry 4.0 use cases such as those found in manufacturing, smart cities, transportation, and healthcare. The potential is arguably unlimited, and CSPs have a huge opportunity to seize the initiative and diversify into the myriad vertical markets that are set to benefit from private or shared 5G networks.
Currently, the emerging use cases are typically focused on private networks and are providing valuable insights as to the key requirements. Rapid setup of private 5G networks is required in order to connect large volumes of devices – cameras, drones, sensors etc., – that are utilized for services needing high capacity and low latency. Open multi-access edge computing (MEC) platforms need to be located on-premise or at the network edge to host MEC applications, as well as the 5G user plane function core, connectivity and security functions. In addition, there must be flexibility to speed up deployments by using hyperscaler edge cloud stacks. Some MEC applications will need intelligent placements in an edge host, others with less demanding service requirements can be hosted regionally – but there will need to be an E2E service orchestration approach in order to automate the service lifecycle.
Getting to the goal of an end-to-end view of the CSPs’ operations is not without its challenges: the adoption of domain automation, sharing of real-time information between domains, implementing real-time inventory to ensure service level agreements (SLAs) are met, and collaboration between and among multi-vendor systems. Real-time inventory is a new essential OSS function that provides an E2E view of the service state and works in conjunction with the assurance and AI/analytics components. CSPs can take a leading role in this complex environment, however the transformation needed is time consuming, complicated, and costly.
CSPs are making good progress with domain automation, although there is still much work to do. It is a new operational approach, combining cloud-native orchestration, inventory, configuration management, assurance and AI/analytics to ensure the domain is fully autonomous. Service design is key – CSPs need to build dynamic service and slice models to populate the service catalog so the orchestration engine can execute requests using service intent. Initial focus should be on pred-defined models for the most popular cross-domain requirements.
End-to-end service management and orchestration provide the critical connection between networks and IT, and the automation this enables is crucial to ensuring the success of 5G network services and the massive potential that they represent.
