BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
The progression of cellular mobile networking is seen all too often as an endless progression of generations, or...
G's. While technical evolution often demands infrastructure redeployment and replacement, that's not always the case. Some of the infrastructure for fifth-generation wireless seems certain to be deployed, but most 5G features that get vendor and media attention are still waiting for a solid business case to provide a return on investment.
Operators are realizing technology advances do not necessarily justify technology transformation. Newer isn't always better, if better is defined as a business case including higher revenue and profits. The challenge for the industry in 5G, software-defined networking (SDN) and network functions virtualization (NFV) is we've proposed technology changes through standards bodies that have a limited scope of consideration and found those limitations mean true business cases can't be built within those bodies. We have to look outside those bodies to address 5G business case issues.
So, it's time to ask whether another operator priority -- 5G service lifecycle automation, also called zero-touch automation -- could help make that business case by making new wireless applications earn a better return on investment.
Beyond 5G's necessary New Radio is the 5G mobile core. The mobile core is a combination of slices of multiple 5G networks that allow multiple virtual networks to be created on top of a shared physical infrastructure. Some or most of these virtual networks might be shared among multiple network owners. This structure is served by a series of higher layers representing different radio technologies, such as satellite, Wi-Fi and cellular, and services, like voice, data, video and the internet of things (IoT).
If 5G core is to advance, it will have to be because it can be justified with revenue and profit capabilities; without service lifecycle automation, that's almost impossible to create. The 5G mobile core framework has been considered a prime place to use NFV and SDN, which has led to broad vendor interest in its development.
Operators are interested, too, but they have two problems with the 5G mobile core. The first is the services that will use 5G core features are either hypothetical or unproven. Many services like IoT, smart buildings or smart cities don't rely on 5G at all -- at least in their early stages. And even these early stage services that aren't 5G-dependent have yet to generate definite revenue opportunities. Without efficient operations, network operators probably can't earn a respectable ROI in mobile applications that use 5G core features.
Operators' second problem with the 5G mobile core is complexity, which is illustrated by previous experience with SDN and NFV. Virtualized infrastructure may be agile, but it often substitutes a series of hosted components and connections for a single device. A hardware appliance that provides for content delivery or serves as an element in an IP Multimedia Subsystem (IMS) or evolved packet core (EPC) configuration may be proprietary and, therefore, expensive. But, operationally, it's simple. Virtual elements aren't simple, so something has to improve operations efficiency, which is where 5G service lifecycle automation would come in.
5G service lifecycle automation could promote efficiency
The structural mixture of 5G layered services and sliced network infrastructure shows why simple doesn't begin to describe the situation. Combine that with a reality where users and services move around geographically and where resources and capabilities are added by operators in a dynamic way, and you have a recipe for complexity. Every mobile operator, satellite provider, Wi-Fi hotspot and content or feature provider will be a new dynamic network resource. Beyond that, the issues include how these capabilities can be incorporated into services.
But 5G service lifecycle management is essential if the versatility promised by the 5G core is to be realized. The vision of a 5G core even redefines the entire notion of services to the point of eliminating the distinction between wireless and wireline -- at least as seen by the user. While the 5G mobile core is made up of layers and slices, new 5G service options will create columns or pillars as vertical associations to those layers. These service relationships will be dynamic, variable in composition and are considerably more complex than static 4G LTE services.
Network slicing that partitions infrastructure is similar to the kind of virtualization NFV offers. Just as with NFV, slicing technology has to be integrated operationally with the rest of the elements of the services being created and supported. The European Telecommunications Standards Institute (ETSI) NFV Industry Specification Group doesn't address this broader 5G service lifecycle automation operations mission of the 5G core, nor do 5G specifications. But without it, many of the benefits of network slicing won't be fully supported, because they won't be managed efficiently.
The layers of infrastructure envisioned for a 5G mobile core are similarly disconnected in a management sense. The model for 5G supports satellite, Wi-Fi, 5G New Radio and even earlier-generation cellular networks. It could be extended to wireline infrastructure. All of these have their own challenges with efficient management to contain operations costs, but what integrates them in 5G?
Integrated management of 5G service needed
Zero-touch automation mechanisms being explored independently in ETSI in the Zero touch network and Service Management Industry Specification Group may or may not address the problem. The group is initially focusing on 5G end-to-end network and service management like network slicing management. The goal is to have all operational processes and tasks executed automatically.
The focus of the group has been on provisioned, per-customer services. The model of 5G primarily addresses activated shared-tenant services, based on already-organized infrastructure and service elements. All of these elastic layers and slices may mandate a completely different vision of what a 5G service lifecycle is, much less how one is managed.
With 5G, operators are forced to consider an expansion of what services fit into a virtual world. In 5G, mobile, content and IoT services are implicitly built on top of service frameworks that create those vertical columns across 5G's slices and layers. The services themselves are more signaled than provisioned, and nobody would expect each phone call to spawn an instance of a 5G IMS or EPC. While the same problem will exist for NFV if it is used to deploy the mobile elements of network slices, it hasn't been addressed. At a minimum, we may need service framework management and something like admission control at the true user-service level.
The 5G challenge in service lifecycle automation is more than adapting to a new mission. It is a challenge of adapting to a whole new vision of what infrastructure is. While 5G still looks at networks as static relationships, the dynamism of software-defined features and software-exploited virtualization is outside that old vision. Service lifecycle management could resolve the problem of making operations efficient in the kind of dynamic infrastructure and services environment that 5G envisions. It could make the difference between a justifiable 5G evolution and a problematic one.