Computer/Embedded Technology

Enabling the 5G intelligent edge takes white-box thinking

30 May 2022 Computer/Embedded Technology

Recent advancements in machine learning (ML) and the internet of things (IoT), paired with new network technologies, 5G New Radio (5G NR), disaggregation, edge computing and open industry standards, are all coming together and leading to a new era that will be driven by the AIoT, or the Artificial Intelligence of Things. Ripples of this transformation will be felt across many industries, impacting business models and operations. Industry 4.0, which is expected to be the next industrial revolution, is one outcome.

AI-enabled private 5G networks will shepherd in an era of connected machines and robotic processes working autonomously to deliver system-wide automated processing control. These systems will have the capacity to ‘learn’ to optimise productivity, allowing businesses to become more efficient and agile. This is expected to significantly improve productivity and output.

In fact, it is estimated that by 2030, the manufacturing sector could add between $400 billion and $650 billion to the global GDP by leveraging the benefits of 5G, and private networks are quickly becoming a new avenue of growth for service providers around the globe. Private LTE and 5G networks generated $4,7 billion in revenue in 2020. That number is expected to be $8 billion by the end of 2023 (according to market intelligence firm, SNS Telecom & IT).

The factory of the future will be more automated and efficient than today, and white-box hardware is an important component in achieving this goal. The concept of white-box hardware has been around for nearly a decade but has really taken off in the networking space in recent years due to the broader adoption of network functions virtualisation (NFV).

A white-box infrastructure is based on interoperable, programmable, off-the-shelf equipment designed to provide support for web-scale, cloud-native, virtualised network architectures – delivering greater flexibility and agility from the core to the edge. This article addresses the reasons, challenges and benefits of service providers and enterprises adopting a white-box edge approach to optimise operations and generate new revenue.

5G – A key enabler for Industry 4.0 and AIoT

Industry 4.0 brings together various advances in technology, but there remains an issue with connectivity. Even the most advanced factories still rely upon traditional fibre-based or Wi-Fi networks. In a fibre-based environment, there is no flexibility for supporting an increasing number of mobile devices and sensors, and in a traditional Wi-Fi environment, the growing number of business-critical connections contributes to interference, congestion and security concerns that are difficult to manage in advanced industrial environments.

When paired with 5G, AIoT has the potential to transform businesses across industries, creating a new economy that will power the fourth industrial revolution; one that is driven by machine learning and automated decision-making, connectivity and the exchange of data. This opens the door to a new generation of smart factories, smart logistics, smart transportation, smart energy grids, smart cities and more.

5G New Radio

In contrast to previous generations, 5G New Radio (5G NR) is built for business. It is a transformative technology that enables greater bandwidth and lower latency. Plus, the added security and reliability benefits meet the needs of mission-critical applications.

Previous mobile communication technologies were designed primarily for mobile handsets. But as millions of new IoT 5G-enabled chipsets, modules and other connected devices come into the market, the focus will go far beyond these personal devices and into a world of connected ‘things’ made smarter through artificial intelligence. A world of AIoT fuelled by 5G.

3GPP has defined three initial 5G use-cases which will help power an AIoT-enabled Industry 4.0:

• Ultra-reliable low-latency communication (URLLC) – for real-time communications between machines.

• Enhanced mobile broadband (eMBB) – greater bandwidth and support for a higher number of connected devices and sensors that generate more data traffic.

• Massive machine-type communications (mMTC) – low-power connections for longer battery life, enabling support for massive numbers of IoT sensors and devices.

Private 5G networks

Private 5G networks have been designed to provide businesses with ultra-reliable, low-latency mobile connections that will power new capabilities – from predictive maintenance, autonomous machines and factory robotics, to ultra high-definition video analytics that will enable many different use-cases across industry verticals.

Today, there are different deployment options and business models for private 5G networks, each with their own level of mobile network operator (MNO) engagement:

Option A: An enterprise builds its own 5G network with spectrum that it has purchased outright or leased from a mobile service provider. Some countries have reserved unlicensed or lightly licensed bands, such as the Citizens Broadband Radio Service (CBRS), to support this capability.

Option B: An MNO builds and operates an on-premises private 5G network for the enterprise and provides integration with its public network if required.

Option C: An MNO builds a public 5G network and provides a dedicated network slice that meets the enterprise’s specific SLA requirements.

White-box Hardware and 5G

With these new business models in mind, service providers are investing in virtualising their mobile networks, creating the agile, automated, cloud-native environment that 5G needs to succeed. An open architecture that utilises common industry standards, supported by the O-RAN Alliance, the Telecom Infra Project (TIP) and others, is needed to fully realise the 5G vision.

Since its foundation in 2018, the O-RAN Alliance has promoted the use of white-box hardware to reduce the cost of 5G deployments. Up until now, mobile networks have been relatively closed ecosystems. By opening the radio access network (RAN) to new cloud and IoT players through a standardised, interoperable, multi-vendor environment, innovative services and applications can be deployed at the edge that generate new revenue streams. Adopting a white-box strategy reinforces this approach – building an open and agile foundation that will allow operators to take advantage of economies of scale and encourage the creation of an innovative and competitive ecosystem that is ready to adopt next-generation cloud-native architectures.

The white-box transformation is also happening at the enterprise edge (Figure 1) with virtual functions such as SD-WAN being deployed on commercial-off-the-shelf appliances or universal customer premises equipment (uCPE), setting the scene to deliver a broader range of premium 5G enterprise services.

As mentioned above, 5G technology enhancements have put the spotlight on private 5G networks as key smart manufacturing and AIoT enablers, providing a secure wireless connectivity solution for AI, VR and AR, which rely heavily on 4K/8K UHD video as a main source of information. NFV and software-defined networking (SDN), along with the definition of O-RAN standards and the use of white-box infrastructure, are democratising mobile network technology, enabling 5G Stand Alone (5G SA), low-footprint solutions that are viable for enterprise deployment.

Still, while building this new open and intelligent 5G edge, a gap needs to be addressed so that the benefits of industrial 5G and AIoT can be fully realised. This is where the operational technology (OT) layer, which is defined as the equipment and industrial systems within a production environment, needs to ‘talk’ to the information technology (IT) and communication technology (CT) layers.

The OT layer needs to integrate seamlessly with the IT and CT layers to deliver on the promise of the intelligent, autonomous, AI-driven factory of the future. Again, white-box solutions can play a key role to accelerate digital transformation and mend this gap by enabling an open architecture that extends network intelligence all the way to the factory floor, over a consistent white-box infrastructure that can reliably execute mission-critical applications anywhere in the network. Private deployed 5G solutions following this approach allow for a lighter, more flexible and future-proof communications infrastructure that can be managed as another enterprise IT system.

White-box benefits

Overall, virtual and open infrastructure solutions based on white-box hardware open the network to innovation, optimise infrastructure costs and lower barriers to entry. The disaggregation of software and hardware is at the foundation of these capabilities. These are the benefits that can be achieved when a carefully considered, strategic approach is taken:

Greater economies-of-scale

The right white-box approach can advance the standardisation of telco infrastructure, making private 5G networks more efficient, flexible and easier to implement with commercial off-the-shelf solutions.

A more agile infrastructure

An intelligent, virtualised network edge infrastructure based on white-box hardware can ensure greater network agility by supporting a cloud-native architecture that enables web-scale open compute, networking and storage.

Hardware interoperability

The interoperability of white-box hardware can provide second-sourcing protection, reducing supply chain risks and enabling a more competitive ecosystem that benefits both vendors and operators.

Challenges at the edge

Today, there is widespread agreement amongst operators that edge computing and core 5G network functions will need to be tightly integrated with and powered by public and private clouds, which is especially true for industrial AIoT. As mentioned previously, the O-RAN Alliance, 3GPP and many operators worldwide understand the advantages and are committed to driving these efforts forward.

But implementing a scale-out white-box strategy across the network is a challenging task as there is no one-size-fits-all option when it comes to the edge. White-box hardware is not all created equally. At its best, it can deliver high availability, redundancy and remote management; supporting intelligent manufacturing by bridging the critical operational technology gap mentioned earlier.

With the arrival of edge computing, the right hardware platform must be chosen based not just on what function it will perform, but also where and under what conditions, to guarantee sustained operations and avoid costly downtime.

Here are some key challenges to consider:

• Reliability – Reliability and sustained operations is a top priority when deploying network equipment that supports mission- and business-critical applications. Painstakingly designed hardware that has built-in redundancy and high availability reduces costly downtime and service interventions. Advanced remote management, such as fail-safe upgrades, is another key consideration that can greatly reduce operational expenses.

• Performance – The data tsunami brought by 5G and AIoT will require higher computing power to be moved closer to subscribers at the edge of the network. In addition, vRAN brings the challenge of deploying baseband virtual Distributed Unit (vDU) functions at the edge on white-box base station infrastructure that will also require higher performance to support 5G NR.

• Energy efficiency – Edge computing moves computing resources outside of the data centre to remote locations with strict energy and space constraints. To ensure maximum energy savings at the edge, performance-per-Watt should be a primary consideration of any white-box strategy, along with hardware acceleration and the ability to maximise density. Hardware consolidation to optimise resource usage will also be critical.

• Environmental – Certain environments, such as heat, cold or dust, can result in poor performance or even failure. As mentioned before, bringing high-performance hardware to remote locations means having to deal with harsh environments far from data centre-controlled conditions. Maintainability is also a consideration. Can the system be easily accessed, maintained and serviced – or are there limitations? It is important to select hardware that is designed to meet these specific edge needs to ensure continued high-performance operations.

• Integration – Network disaggregation and virtualisation opens the market and optimises infrastructure costs. However, this new multi-vendor scenario creates new risks in having to integrate pieces from different sources without jeopardising performance, security or reliability. An open, programmable, off-the-shelf white-box approach requires strong ecosystem collaboration and pre-validated configurations to achieve streamlined deployments. In the end, this saves time and money while enabling more agile, future-proof, open solutions.

The Advantech advantage

Selecting the right 5G white-box approach is critical to avoid unpleasant surprises when the rubber meets the road. Advantech’s 5G edge servers and appliances provide a solid and open foundation for service providers and enterprises to deploy agile, reliable and secure network edge solutions that break away from monolithic architectures.

These scalable white-boxes integrate the latest computing and networking technologies into optimised edge platforms that are widely deployed, running popular on-premises, access network and edge cloud software from industry-leading ecosystem partners. This flexible approach helps minimise supply chain risks. It protects network investments, building an open and virtual infrastructure ready to adopt next-generation converged and cloud-native models.

Advantech 5G edge solutions integrate lessons learned from over 20 years of experience designing carrier-grade hardware for leading service providers and telecom equipment manufacturers globally. Some of these include:

• High performance and density: Unlike IT servers, Advantech 5G edge servers are designed from the ground up to optimise throughput and offload resources required by communication workloads. The systems not only combine powerful CPUs with support for high thermal design power (TDP) PCIe cards, but also carefully balance I/O between multiple processor sockets (NUMA balanced).

• Built-in reliability and security at the edge: Advantech’s advanced designs yield higher margins and lower component stress for improved platform reliability. The servers support single failures of critical components such as power supply modules and fans. They have been designed for the telecom edge, including NEBS Level 3 compliant and ultra-short depth (11”) configurations that pay special attention to maintainability and serviceability. In addition, redundant BIOS and firmware images not only provide a safe way to recover from component failures but also offer remote fail-safe update capabilities via Advantechs IPMI, which reduces MTTR and costly on-site services.

• Flexible integration, customisation and design: Solution providers can also leverage Advantech’s customised COTS framework for semi-custom electronic or mechanical designs, as well as product branding including artwork, packaging and BIOS firmware strings or IDs. As Advantech designs and manufactures all its sub-assemblies, it can modify and optimise any element in the system to suit a specific market need.

• Full ecosystem collaboration: Advantech works closely with leading silicon, OS, ISV, system integration and service provider partners to jointly address the challenges of open and disaggregated networks, bringing to market optimised solutions that have been verified to perform well together.

Building the backbone of the new AIoT economy

5G edge servers for the open RAN

Advantech SKY-8000 servers are high-performance, high-density platforms optimised to run virtualised communication workloads at the network edge.

Deployed globally in telecom and industrial business, and mission-critical applications, they provide a wide choice of white-box hardware platforms that scale from pole-mount, ultra-short-depth, extended operating temperature range, outdoor configurations all the way up to high-performance designs that can be deployed as micro data centres at the edge or in virtual central offices.

Universal edge appliances for SD-WAN, uCPE and SASE

Advantech FWA network appliances streamline enterprise network transformation with universal platforms that can scale secure connectivity services from small and medium branches to large campuses and headquarters. Rugged configurations such as the FWA-1112VC, supporting extended operating temperature ranges and 5G, bridge the gap between enterprise and industrial connectivity.

4K/8K video and AI acceleration

Advantech provides commercial off-the-shelf access to live video acceleration technology in platforms that have been designed for edge deployment.

Advantech VEGA video solutions enable highly efficient, real-time 4K/8K compression and streaming in compact modules and appliances.

Private 5G network small cell demonstrator

Advantech wants to streamline solution evaluation and validation by providing enterprises with a fully integrated small cell demonstrator that can be easily deployed on premises, delivering all required elements to run private 5G NR networks in a ruggedised rack. The demonstrator can be configured

to meet different performance and functionality needs in indoor or outdoor scenarios.

Coupled with Advantech solutions for network security, SD-WAN, 4K/8K and AI, this demonstrator can help enterprises accelerate their digital transformation toward true OT, IT and CT integration to achieve operational excellence.


Share this article:
Share via emailShare via LinkedInPrint this page

Further reading:

UFS Ver. 4.0 embedded Flash memory devices
EBV Electrolink Computer/Embedded Technology
KIOXIA Europe has announced sampling of the industry’s first Universal Flash Storage (UFS) version 4.0 embedded Flash memory devices designed for automotive applications.

Powering factory automation into the future
Rugged Interconnect Technologies Computer/Embedded Technology
Powered by the newest 13th Gen Intel processors, ADLINK Technology’s COM-HPC-cRLS module is a future-proof edge AI solution.

Linux OS with immutable file system
Computer/Embedded Technology
What really sets VanillaOS apart from others, however, is in security, where it takes a new approach to computing by using an immutable file system for improved security and stability.

Rugged laptops support the renewable energy sector
Osiris Technical Systems Computer/Embedded Technology
The Durabook S15AB semi-rugged laptop was recommended due to its military-grade durability, powerful functionality and desktop performance, allowing it to keep pace even for intensive tasks such as asset management and maintenance.

Memory comparison: DDR4, DDR5, GDDR6 and HBM
Avnet Abacus Computer/Embedded Technology
DDR5 SDRAM was introduced in 2021 and brings many performance enhancements, but just how much faster is it over DDR4? And how does it compare to other memory types like HBM2e and GDDR6?

Industrial networking devices from InHand
Otto Wireless Solutions Computer/Embedded Technology
The InHand routers, edge gateways, and industrial switches are ideal solutions for any scenario in which control of an IIoT Ethernet network is needed.

Availability of a wireless network
RJ Connect Computer/Embedded Technology
The AWK-3252A Series 3-in-1 industrial wireless AP/bridge/client from Moxa is designed to meet the growing need for faster data transmission speeds through IEEE 802.11ac technology, for aggregated data rates of up to 1,267 Gbps.

IA-820i with PCIe Gen5 interface
Avnet Abacus Computer/Embedded Technology
Featuring the Intel Agilex 7 I-Series FPGAs, these cards are optimised for applications that are bandwidth intensive and require high-performance processor interfaces.

Hi-Rel for IIoT applications
Rugged Interconnect Technologies Computer/Embedded Technology
In the increasingly complex world of industrial applications, having industrial-grade solid-state drives (SSDs) with excellent performance and high endurance can ensure stable operation of equipment.

Compact edge computing system
Altron Arrow Computer/Embedded Technology
The Boxer-8621AI is equipped with the superior edge computing performance of the NVIDIA Jetson Orin Nano, to provide the power to execute all AI applications.