The PowerPoint presentations are compelling. On the face of it, network slicing enables the agility needed for 5G to fulfil the promise of service-centric network architectures. Slicing promises to introduce network agility for operators to efficiently differentiate their network services and reinvigorate their market opportunities.
Mobile traffic is growing 40-50% per annum in some markets, and CSPs need to continue to invest to meet that capacity. Theoretically, 5G means increased network capacity in the region of ten times. In other words, 5G allows CSPs to provide more capacity at a lower cost, and that is the heart of the 5G business case. Sound familiar? It should. That is the same business case that drove CSPs to upgrade to 4G.
Network slicing unlike conventional network resource management schemes, which are largely static, network slicing enables dynamic resource allocation capabilities that simultaneously respond to the demands of multiple services. For example, a sliced network might simultaneously support conventional, mission critical and latency sensitive services on the same network, with dynamic resource allocation mechanisms to ensure that each service is managed according to its real time capacity and service quality demands.
See Also
5G Insights November 2018
Three steps to monetizing 5G
Barriers to adoption for 5G-capable smartphones
5G mobile backhaul isn't all fiber
While slicing as a concept is straight-forward, implementing slicing is complicated.
Slicing requires advanced cloud, network virtualization and edge computing functionality with architectures that simultaneously support diverse service demands on the same network fabric. Slicing depends on programmable architectures that can dynamically manage and orchestrate network resources and traffic-flows on an end-to-end basis.
Solutions must traverse heterogeneous radio, transport and core network environments and cope with multi-vendor and multi-generational environments. Furthermore, some service demands, such as ultra-low latency connectivity, require network architectural changes to include targeted functionality, such as edge computing.
Pragmatic strategies
As operators implement network slicing, they require measured and pragmatic strategies to realize the capabilities that slicing enables. These strategies depend on continued advancements in cloud and network virtualization to enable end-to-end network programmability, and service management and orchestration (MANO) and business process automation to respond to the salient characteristics of the services being offered.
When operators implement slicing, the network MANO demands will grow astronomically and are likely to create a tangible means for operators to differentiate their services. Slicing will overwhelm conventional MANO architectures, even those being developed for today’s SDN and NFV architectures. This is particularly the case as operators scale the number of end-to-end slices that are simultaneously supported on their networks. Some operators have stated that they plan to have hundreds or possibly thousands of network slices to support their future services. To achieve this, MANO solutions would require the support of advanced artificial intelligence and machine learning techniques that have yet to be developed.
Today operators generally have business processes and go-to-market strategies that are designed for mass market service offers with relatively basic commercial frameworks. This has been a primary contributor for network service commoditization, particularly with the growing market demand of digital services. While slicing aims to stem the tide of network commoditization, it depends on new business processes and support systems to reliably commercialize the diverse and dynamic services that slicing enables. For example, when operators use slicing to implement tiered services it is likely that they will require real time charging and settlement solutions to accurately identify, track, monetize and settle multitudes of slicing events, when they occur.
Complementary to 4G
Even though 5G and network slicing aims to disrupt traditional networks and services, 5G will generally be deployed as a complementary solution to 4G for the foreseeable future. Operators are well positioned to capitalize on current and emerging 4G technology and are not necessarily deploying 5G with the same ubiquitous coverage.
This is particularly the case of 5G-mmWave radio technology, which offers tremendous peak data rates but with sparse and unreliable coverage characteristics. For 5G-mmWave, operators must anticipate (as best they can) its coverage availability and ensure that their network slices and associated service plans are optimized accordingly.
In addition, some enterprise applications, such as Industry 4.0 manufacturing, are campus based and are likely to only require local 5G coverage with edge computing capabilities. For localized enterprise applications, particularly where on-premise equipment is used, slicing might be unnecessary, or if implemented have relatively static resource allocation architectures that is optimized for the specific service demands of the enterprise.
Even though network slicing has been a core value proposition since the inception of 5G, slicing is still nascent and requires significant technology advancements to ensure that it operates reliably with mass market scale. Operators require new commercial relationships, business processes and network and service delivery architectures to monetize network slicing. Furthermore, continued advancements in network virtualization, service MANO and network resource optimization capabilities are crucial for network slicing to function effectively. If implemented correctly and treated as an integral component, rather than merely a feature of 5G, network slicing can be a game changer for operators as they seek opportunities to increase the intrinsic value of their networks.
Phil Marshall is chief research officer of Tolaga Research
This article first appeared in Telecom Asia 5G Insights November 2018