In recent WILTWs, we have argued that the buildout of 5G wireless networks is set to be one of the largest infrastructure projects ever and is the most important technology race of the next five years. We are studying the implications of 5G wireless as it will likely determine the winners of the 21st century (see reports).
Fundamentally, in a 5G world, networks are shifting from being hardware centric to software driven. Automation is critical as it will enable 5G networks to cost effectively and efficiently handle the exponential increase in the number of connected devices and thousand-fold jump in data generation produced via the IoT, autonomous cars, and industrial applications.
Conventional network architecture alone will not achieve 5G’s promise of low-latency, reliability, and dynamic service applications. Intelligent network systems will be vital in operationalizing and scaling 5G services.
5G networks need next-generation automation capabilities to be deployed in parallel, if not in advance of the physical infrastructure build-out to achieve its full-potential. Leveraging network automation, predictive analytics, and AI algorithms are key to service provider near-term profitability and essential to long-term survival, notes a Strategy Analytics white paper.
Network automation can cut operational expenses in half over a three-year period for operators, concludes a recent MIT Technology Review study. Automation will save energy and backhaul expense, reduce engineering costs per gigabyte, and lower the cost of managing cell-sites. AT&T states that it will achieve 40% to 50% in OPEX savings, once three-quarters of its network is automated (expected in 2020). Network automation will also improve spectrum utilization and optimize cell-site capacity resulting in capital-expenditure-efficiency advantages.
What does it mean for investors?
- New networking technologies and architectures are fundamental to automating 5G. Networks will become increasingly complex to support 5G-enabled services.
The transformation of operators’ networks to a web-based application programming interface (APIs) will facilitate agility, support faster-experimentation, proactive error detection, and automated resolution.
- 5G networks will be centered on “Service-Based Architecture” (SBA). SBA supports the reuse of software components and interconnections, allowing faster rollout of new applications, notes Ericsson. The SBA-enabled 5G network cloud would function similarly to cloud-computing based infrastructure-as-a-service (IaaS) model, transforming communication service providers (CSPs) into application innovators for vertical industries, notes Analysys Mason.
Network-functions virtualization (NFV) and software-defined networking (SDN) are the first building blocks of flexible and programmable automated networks. NFV helps achieve higher network efficiency through optimum bandwidth utilization. NFV also allows multiple virtual networks on a single physical infrastructure reducing the need of hardware and time to install.
SDN provides an intelligent architecture for network programming, on-demand resource allocation, cloud security, and helps create new services on top of virtualized resources. Network slicing leverages NFV and SDN to customize on-demand bandwidth allocation ensuring quality of services for mission-critical 5G-use cases.
The Open Network Automation Platform (ONAP) will also be vital in providing vendor-neutral platforms to operators—enabling cable and cloud companies to design, implement, and manage differentiated custom services. This will accelerate cross-domain automation and innovation.
- Digitization of operators will ultimately lead to fully-autonomous “zero-touch” networks. The European Telecommunications Standards Institute (ETSI) is building on NFV and SDN to provide a fully-automated end-to-end network architecture solution.
Last year, ETSI led Zero-touch network and Service Management Industry Specification Group (ZSM ISG) launched ServoCloud, a first prototype to automate operations starting with 5G-network slicing. AI/ML integration into automated networks will also minimize human intervention by delegating complex and mundane tasks to machines, reducing error rates and cost.
- Operators have already started building in-house teams and implementing technologies for automation. AT&T, in collaboration with Tech Mahindra has developed “Acumos AI,” an open-source framework that makes it easy to link multiple micro-services together. Verizon recently tested multi-access edge-computing equipment (MEC) and platform software on a live 5G network in Houston that resulted in 50% reduction in latency.
- Numerous studies suggest the network-automation-market size could reach tens-of-billions of dollars, led by NFV and SDN technologies. Digital infrastructure and network automation will drive total telecom-software expenditures to reach $95.1 billion in 2022, up from $75 billion in 2017, notes a study by Analysys Mason. A Global Market Insights report estimates that the NFV market will exceed $70 billion by 2024, led by software implementation. The SDN market is expected to grow at a CAGR of over 42% to reach $59 billion by 2023, notes research by Market Research Future.
- Compelling candidates for a network-automation portfolio include:
- Ciena Corp. (CIEN, $44.06) – offers orchestration, analytics, and network-slicing control software for virtualized 5G network infrastructure—trading at 10.9 times consensus FY20 EV/EBITDA, below its trailing average of 12.1 times.
- Juniper Networks (JNPR, $26.85) – provides automated networking systems with open frameworks, APIs, and toolkits—trading at 15.1 times free cash flow and 7.9 times consensus FY20 EV/EBITDA, below its trailing median of 9.1 times.
- Ericsson (ERICB SS, 93.00 SEK) – is a market leader in NFV and network orchestration—trading at 7.8 times consensus FY20 EV/EBITDA, below its trailing median of 11.0 times.
- Cisco (CSCO, $56.13) – offers its Crosswork Network Automation solution, which delivers open APIs to accelerate application and service development—trading at 16.8 times free cash flow.
- VMWare (VMW, $173.42) – offers a suite of automated solutions for programmable and secure software-defined networks—trading at 19.8 times free cash flow and 16.8 times EV/EBITDA, below its trailing median of 18.4 times.
- ZTE (763 HK, 21.75 HKD) – offers intelligent network solutions, which shortens the time-to-market from hours to minutes, improves O&M efficiency by more than 25%, and reduces OPEX by more than 50%–trading at 12.1 times consensus FY20 EV/EBITDA, below its trailing median of 15.2 times.
- Nokia (NOKIA FH, 4.46 EUR) – offers applications to automate services across multiple network layers—yielding 4.5% and trading at 6.8 times consensus FY20 EV/EBITDA.