Six Threats & Opportunities of the Disruption Called 5G

I grew up in the wireless networking business before switching to Automotive and Industrial markets. As discussed in my previous blog (Car Meets Cloud), the future of Automotive/AV & IIoT is intertwined with wireless networking and specifically 5G, which makes me excited about turning the full circle.

In this blog, I will take a closer look at 5G and the topic of Multi-access Edge Computing (MEC) vs. Cloud Computing (which we started in the last blog) and the opportunities and implications on the Cloud. We will see how cloud and communication networks' convergence allows network operators to roll out 5G networks more efficiently and cost-effectively. This will provide a massive opportunity for large Cloud infrastructure providers such as Microsoft, AWS, and Google. 5G will create new markets. Just as "software is eating the world" in other technology spaces such as Automotive, new players with software expertise will enter the market and challenge the incumbents in the networking space.

My six (6) observations will show you the scope of the threats to be aware of and the opportunities ahead. Success is where preparation and opportunity meet, so it’s time to get prepared.

5G – Not all Gs are created Equal

5G wireless technology is the 5th generation mobile network. 5G promises to connect everyone and everything, including machines, objects, and devices. Based on my career history, my favorite feature is that 5G will be the first end-to-end fully software-defined architecture. 5G enables wide use of cloud virtualization technologies such as network functions virtualization (NFV) and software-defined networking (SDN). 

Here, I am briefly introducing the different channel access schemes, not to get into the signal processing requirements of 5G, but rather to show later on why a software-based implementation leveraging a particular class of semiconductor technology is a great way to go.

While 2G systems used Time Division Multiple Access (TDMA) channel access schemes, 3G used a different method called Code Division Multiple Access (CDMA) to share each channel. 4G introduced a scheme called Orthogonal Frequency-Division Multiplexing (OFDM), which in a way was a 2D access scheme with modulation in frequency using sub-carriers and the use of time slots for users to take turns and share the channel. When it comes to channel access, 5G is a lot more complex and employs massive MIMO (multiple inputs multiple outputs) antenna technology using beamforming. 5G allows antennas to be small, enabling immense arrays of antenna elements in a reasonably sized physical space. Beamforming is the ability to adapt the radiation pattern of the antenna array to a particular scenario.

5G signal processing with hyper-MIMO and beamforming requires a tremendous volume of data to be processed. This involves many matrix operations and vector processing, which are best performed using Graphical Processing Units (GPUs) and Tensor processors for fast parallel computations. For example, in 5G base stations, among other things, multi-user decoding schemes deal with large matrix multiplications that are performed very efficiently using GPUs.

5G: Driving Trends in Networking

Trends in networking consist of

·        Decomposition of software and hardware

·        Control and User Plane Separation (CUPS)

·        â€œSoftware-Defined” leveraging NFV & SDN efficiencies

·        Support for 5G Radio (NR) and dual-connectivity handsets

·        Use of off-the-shelf technology

·        Virtualization

·        Decentralization, Distribution, and Disaggregation

·        Open APIs

·        Programmable based on real-world network conditions

·        Flexible deployment architecture

·        Support for public and private clouds

·        Interface interoperability and support for multivendor deployments

Virtualizing the network through the use of cloud-native models in a service-based architectural framework is an integral aspect of advances in the 5G system. 5G core is expected to be access-neutral, integrating fixed and mobile, terrestrial, and non-terrestrial, local- or wide-area access networks.

The above trends introduce drastic disruption to the telecom market. This is when we have seen name-brand traditional players losing share, and newcomers have moved up the ranks in other markets. Below are the market share estimates for 5G base stations from TrendForce.

Source: TrendForce

Based on the geopolitical issues between Washington and Beijing, Huawei and ZTE shares will erode in the international markets. The domestic Chinese market is large enough, however, to propel their growth. The aggressive 5G deployment in South Korea is benefiting Samsung. The interesting group of challengers residing in the "Other" category (the O-RAN community) is awaiting technology maturity, and to fill the void from the absence of Huawei and ZTE from the international scene.

5G will support a host of services and devices and compliant to a vast set of requirements around capacity, coverage, reliability, throughput, and latency. 5G must support a range of data rates, from small sensors with a low-data rate of 10 kbps to new immersive high-bandwidth/high data-rate experiences in highspeed moving transportations at multi-Gbps. At the heart of 5G is the new 5G New Radio (NR) unified air interface designed to create a unifying connectivity fabric for society's needs for the next decade and beyond.

5G’s benefits vs. 4G include:

·        3x spectrum efficiency – leveraging advanced antenna techniques

·        100x traffic capacity – leveraging more small cells everywhere

·        100x network efficiency – optimizing energy consumption with more efficient processing

·        10x experienced throughput – bringing more uniform, multi-Gbps peak rates

·        10x decrease in latency – as low as 1 ms

·        10x connection density – efficient signaling for IoT connectivity

·        10+ years of battery life for IoT sensors

·        Software-Defined Virtual Networks

·        A new paradigm for flexible service offerings called network slicing·        

·        Easier migration from 4G

Source: Qualcomm

5G allows for three primary use cases that may be tuned to the particular requirements of their customers' business models:

·        eMBB: Enhanced (or extreme) Mobile Broadband - Video-centric applications that consume a lot of bandwidth and will be the source of the most traffic on the mobile network. It also enables broadband services everywhere through fixed wireless access (FWA), delivering speeds comparable to the current fiber-based system. 

·        mMTC: Massive Machine Type Communication (IoT) – Much larger-scale version of today's IoT applications, with billions of devices connected to the network. Devices will generate much less traffic than eMBB, but orders of magnitudes more of them.

·        uRLLC: ultra-Reliable Low Latency Communication – Applications requiring "real-time" control, such as automotive V2X or remote surgery, need Mobile Network Operators (MNOs) to have mobile edge computing in place. This use case post-release 16 will be denoted as eURLLC.

Source: Qualcomm

Release 16 (July 2020) enhances C-V2X’s Sidelink by expanding it to newer applications such as public safety, emergency services, and others and further improves its performance. One may say that the new features would seal the fate of DSRC.

Release 16 also brought a long list of foundational enhancements in terms of capacity, coverage, latency, power, reliability, mobility, ease of deployment, and more. The two areas that I am most excited about are the enhancements to MIMO performance (listed in the graphics below from Qualcomm), and enhanced ultra-Reliable, Low-Latency Communication (eURLLC), hence the addition of the “e” upfront.

Source: Qualcomm

To enhance the uRLLC use cases for factory automation and automotive, release 16 also improves the link reliability (up to 99.9999%).

According to McKinsey, adopting the above 5G use cases will "likely happen in waves" in major markets. eMBB will reach mass-market adoption first, followed by eURLLC, and lastly, mMTC. The speed of adoption and rollout of the three use cases will depend on the rate and coverage of 5G network deployments in a given municipality, the evolution of laws and regulations specific to a given country, and the availability of appropriate 5G chipsets, and ultimately the relevant devices to go along with them.

Source: McKinsey

Based on the McKinsey analysis here, there will be some time before the growth for each use case will get accelerated. Based on the graph here, the eURLLC use case volumes will have to wait until L5 autonomy takes off, which will be after 2030. Besides that, there are use cases such as remote robotic surgery that can drive moderate volumes before 2030. The volumes favor IIoT type of applications related to mMTC use cases. The volumes, however, do not necessarily indicate which applications will be the most profitable.

As we have discussed, 5G brings obvious advantages and enables many services within each use case. The technology advantage, however, needs to create economic value for the given services to take off. If it's not medium-term and long-term profitable, it will not become self-sustaining and will die away. There are B2C and B2B advantages that 5G brings to bear. Here I am only considering the B2B use cases. For B2B, McKinsey’s analysis points at 5G use cases in the following industry segments:

·        Industry 4.0, including autonomous systems in factories

·        Smart cities, with applications such as HD cameras to monitor the safety

·        Smart energy, such as intelligent grid control

·        Connected offices, including sensor-based building management

·        Smart security, including the provision of emergency services

·        Connected health, such as mobile medical monitoring

While today we don’t have visibility to other use cases, it doesn’t mean that new applications that may emerge later on won’t be around the corner.

The three primary use cases above are supported by network slicing, with each slice having its architecture, management, and security. 

What is Network Slicing?

Since we will be talking about RAN virtualization, before we go further let me clarify the concepts of Open RAN, vs. O-RAN, vs. vRAN. While there could be some overlap, they are not the same, even though they may sound the same.

Open RAN - Open RAN, which applies to ALL Gs (2G thru 5G), is a movement in wireless telecom supporting vendor neutrality and the use of general-purpose commercial off-the-shelf (COTS) HW platforms, and the disaggregation of the hardware (HW) and software (SW) functionality. This allows one vendor to provide the RAN HW, another providing Cloud Infrastructure, and another developing SW. In this case, the RAN functionality is built using open interface specifications between elements that could be from the O-RAN Alliance. This is different from purpose-built RAN products by a single network equipment provider. As an example, Ericsson’s Cloud RAN product follows the Open RAN topology.

O-RAN: O-RAN (with the hyphen) is about the O-RAN Alliance specifications, which takes this concept multiple steps further with many open interfaces from different vendors. O-RAN Alliance is a specification group that publishes new RAN specifications releasing open software for the RAN and supports its members in integrating and testing their products.

There is also the concept of vRAN (or Virtualized RAN), which is an implementation of the RAN, where it virtualizes non-real-time function of the RAN (namely DU & CU) to run as cloud-native functions leveraging containers, using general-purpose processors on any cloud. In ALL cases, the overall solution needs to be fully system verified, and one entity stands behind the overall solution (one throat to choke by the customer) or the benefits of open standards will not come to fruition.

Network slicing enables the creation of multiple virtual networks from a single physical network. With cloud virtualization technologies such as network functions virtualization (NFV) and software-defined networking (SDN), resources and functional components may be shared across network slices. Network slicing partitions the core network into multiple virtual networks supporting different Radio Access Networks (RANs). Each virtual network includes a set of logical network functions isolated from one another. Slicing the network provides a unique set of optimized resources and network topology based on customer requirements. These are arranged with their network architecture and provisioning, providing self-governing orchestration and supervision. Simultaneously, features such as throughput, data speed, latency, quality of service, reliability, and connectivity can be customized in each slice, based on specific Service Level Agreements (SLAs).

The benefits of network slicing are:

·        It allows for differentiated services that people will pay for to roll out quickly.

·        It allows for more effective management of network traffic and costs. 

The challenge of network slicing is that this is a new concept for Mobile Network Operators (MNOs), who are very conservative and cautiously rolling out new features, with limited pricing models.

Source: GSMA

Chasing Software Defined Radios (SDRs) and Networks (SDNs)

Earlier in my career, I was a telecom solution architect at Cadence Design Services, where a significant source of our revenue came from base station channel-card cost reduction. As wireless standards were continually evolving, and most solutions were "hardwired," no engineer wanted to work on older generation products. At the time, software-based approaches were not cost and power-efficient, and the industry's mindset was different. A couple of years later, after reading a handful of white papers about Software Defined Radios (SDRs), I joined a pioneering start-up called Morphics Technology. I then evangelized a software-programmable hardware-reconfigurable (micro-coded) engine for wideband data-flow and multi-standard wireless communications. At Morphics, we boasted of having the first true Software Defined Radio (SDR) in the industry, and coined the phrase "Node-B on a Chip" for our 3G WCDMA Node-B processor. Unfortunately, the market wasn't ready for SDRs, and Morphics failed. A couple of years later, following my SDR/SDN passion, I joined the Motorola networking division, chasing the same dream of developing a true SDR-based baseband channel card. I resurrected the name “Node-B on a Chip” and using Motorola’s micro-coded engine test-chip (RCF) and StarCore DSPs, my team and I developed TD-SCDMA and WCDMA baseband channel cards. We had moderate success in China and Europe, but the market still wasn’t ready. By 2009, when at Freescale, I helped our new networking team working on a new Femto/Micro/Pico platform to form a partnership for their L1 software and baseband channel-card reference designs. The market seemed a lot more receptive to SDN-based approaches. Now with 5G, the architectural foundation will be fully end-to-end software-defined.

At both Cadence and Morphics, I worked with my friend Joe Boccuzzi, who has also been a huge fan of SDRs and SDNs since the early 2000s. He is now at NVIDIA working on what else – 5G Virtual Radio Access Networks (vRAN), in a fully software-defined solution.

NVIDIA’s Solution

5G network architecture needs to be fully software-defined, dynamically reconfigurable, and easily deployed and managed. A 5G, Cloud-native Radio Access Network (C-RAN) is a software-defined computing architecture that brings real-time, high-bandwidth, low-latency access to 5G communications. C-RANs are ideal for both centralized and distributed RAN architectures.

NVIDIA’s key components of an end-to-end 5G C-RAN edge computing system include NVIDIA GPUs, NVIDIA Mellanox Smart Network Interface Cards (NICs), cloud-native system software (Kubernetes and containers), and software components such as the NVIDIA Aerial SDK built on CUDA SW.

Earlier in the blog, I mentioned that 5G signal processing with hyper-MIMO and beamforming is complex, with a massive volume of data to be processed. This involves many matrix operations and vector processing, which are well suited to be performed using Graphical Processing Units (GPUs) and Tensor processors for fast parallel computations. NVIDIA is the GPU industry leader with ~75% market share, so it shouldn't surprise that it would leverage its CUDA GPUs for baseband processing and Virtual Network Function (VNF). The Aerial SDK implements L1 PHY processing leveraging cuVNF and cuPHY SDK. Virtual Network Functions (VNFs) are virtualized network services running on open computing platforms formerly carried out by dedicated proprietary hardware technology. Common VNFs include firewalls, WAN optimization, virtualized routers, deep packet inspection, and network address translation (NAT) services. VNFs can be linked together in a process known as service chaining.

The cuVNF SDK (CUDA-based VNFs) provides networking libraries and features to optimize packet placement and data transmission and reception, to and from the GPU memory, supporting traditional VNFs as outlined above.

Source: NVIDIA

The NVIDIA cuBB SDK provides all of the 5G Layer-1 (L1) signal processing, with best-in-class throughput by keeping all physical layer processing within the GPU’s high-performance memory pipeline. It also uses a scalable single software programming model. Aerial 5G cuBB (CUDA Baseband) provides transport packets to the third-party L2+ stack. The L2+ stack performs the MAC, RLC, and PDCP functionality that interface to the core network (EPC). The present Aerial SDK supports up to 16 downlink MIMO layers and eight uplink MIMO layers. As each layer is basically a parallel transmission channel transmitted simultaneously utilizing the same resources, spectral efficiency increases.

Traditional systems use a CPU host memory and static-function hardware to accelerate portions of the 5G physical layer pipeline. This look-aside architecture causes repeated transfers of data in and out of local CPU caches, causing bottlenecks across bandwidth-constrained PCIe (PCI express) buses and system-level performance degradation. Now comparing the inline architecture, GPU-Direct Remote Direct Memory Access (RDMA) is used to save computation cycles for memory access. This allows the physical layer data to remain within the GPU's high-performance memory subsystem, keeping the GPU processing engines operating at peak efficiency. cuVNF provides the header/data split functionality that allows agile packet filtering. O-RAN flow identification allows faster packet flow steering based on the MAC address.

Using a C-RAN implementation on GPUs with CUDA programming results in faster compute capability for matrix calculations for the 5G PHY layer functions.

Source: NVIDIA

The solution provides flexibility and scalability to enable network slicing services in a software-defined solution, supporting all 5G use cases efficiently.

The overall solution also includes Mellanox’s ConnectX-6 Dx SmartNIC (Network Interface Card) product. A NIC card is a PCIe card that plugs into a server or storage box to enable connectivity to an Ethernet network. SmartNIC products provide the basic functionality of a NIC, with added features like precision timestamping, DPDK (Data Plane Development Kit). A DPU (Data Processing Unit) based SmartNIC adds the benefits of a DPU co-processor, implementing network traffic processing on the NIC typically performed by the CPU, hence offloading its resources.

NVIDIA’s use of 5T for 5G technology embedded in Mellanox ConnectX-6 Dx SmartNIC product ensures clock accuracy of 16ns or less, which exceeds stringent industry-standard timing specifications for eCPRI-based RANs. 5T-for-5G (Time-Triggered Transmission Technology for Telco) implements the IEEE 1588v2 PTP clock in the SmartNIC. The IEEE 1588v2 standard defines the Precision Time Protocol (PTP), which synchronizes clocks throughout a network. The standard describes the PTP boundary clock's hierarchical master/slave architecture for the distribution of time-of-day. 5T for 5G delivers accurate timing synchronization across front-haul and mid-haul networks.

NVIDIA’s Aerial SDK is Kubernetes-based and supports container-orchestration for ease of container deployment and management. It supports AWS Cloud and Microsoft Azure as well as NVIDIA’s own EGX platform.

Migration from 4G to 5G

Learning from previous transitions and in anticipation for the transition to 5G networks, the Telecom industry has made great progress re-architecting 4G LTE evolved packet core (EPC) also into a Control and User Plane Separation (CUPS), and 5G User Plane Function (UPF). CUPS play an essential role in Multi-access Edge Computing (MEC), particularly in the case of enhanced Ultra-Reliable Low-Latency Communications (eURLLC). For example, to support autonomous vehicles, a Mobile Network Operator (MNO) can quickly scale up their user plane capacity at the edge. By adding new gateway capabilities, MNO can decide which traffic could remain local and which needed to be backhauled through the core network. This ultimately reduces network traffic in high-traffic areas, ensuring that vehicles continue to receive reliable, low-latency, and real-time information.

Like 4G, 5G will not happen overnight, and mainstream adoption is likely to be a few years away. However, 5G will evolve while co-existing side by side with 4G. Carrier Aggregation, Dynamic Spectrum Sharing, and Dual Connectivity are all critical to make coexistence work. This is important, as it allows a more gradual investment requirement for mobile network operators (MNOs). If we separate the radio section from the core network, we can define the gradual transition from 4G to 5G and the associated investment. Eventually, MNOs will build standalone 5G networks from scratch.

5G uses a New Radio (NR) for a more capable 5G wireless radio interface, whereas 4G uses an LTE radio. The graph below shows the topology of using the 5G NR with Evolved Packet Core (EPC), the framework being used for 4G Long-Term Evolution (LTE) network, for providing converged voice and data. The EPC is connected to the external networks and may include the IP Multimedia Core Network Subsystem (IMS).

Today most 5G deployments are based on NR non-standalone (NSA) technology, which uses existing LTE radio access for signaling and EPC networks enhanced to support 5G NSA. This allows a quick rollout of 5G services while maximizing the reuse of existing 4G networks and a more gradual investment requirement by the MNOs.

To truly unlock 5G’s full potential and deliver new 5G differentiating services, a Service-Based Architecture (SBA) needs to be rolled out based on an NR standalone (SA) and 5G Core (5GC) configuration. This is where it can get interesting for Cloud IaaS and PaaS providers, as SBA will be built using IT network principles and cloud-native technology. This means 5GC will be implemented using microservices, containers, centralized orchestration, CI/CD, open APIs, service meshes, etc. Some MNOs and Telcos have already virtualized different parts of their network, but now they need to virtualize all parts of their network, and moreover “cloudify” their networks.

5G is an inflection point. It is a crucial driver for digital transformation and will become the connectivity fabric between people, things, systems, workloads, and business processes, resulting in a new productivity era. But like any other inflection point, there are both threats as well as opportunities that members of the ecosystem need to deal with.

Threats and Opportunities Coming Out of the Disruption Called 5G

The disruption caused by the rollout of 5G networks will positively or negatively impact many players, including the MNOs and Telcos (Communication Service Providers) themselves, depending on how they play the game.

Time For MNOs and Telcos to Change Course

The digital market’s growth has added hundreds of millions of younger customers for MNOs and Telcos; however, it has also shrunk the profit pool due to Over-The-Top (OTT) service providers. The digital native OTT players such as Apple’s FaceTime, Skype, Google Hangouts, Zoom, Tencent QQ, or Tencent’s WeChat, and many others offer voice, messaging, and video calls used to be the domain of traditional Telcos. The Covid-19 lockdowns and new work at home policies have exasperated the situation. With the always-online generation’s help, OTT players now own over 80% of all messaging traffic and account for a significant portion of international voice traffic minutes. This has, in effect, turned MNOs and Telcos into “dump Pipes.” All that heavy CAPEX and OPEX, only for the revenue per bit coming down, and the profits to go to the OTT players. This has led to a sharp dip in traditional, cost-heavy analog voice and text services and strong growth in usage of services provided over the web. While the demand for telecommunications services globally is at its highest level, MNOs and Telcos’ financial challenges have become worse, as OTT players have taken a considerable chunk of telecom’s market share, decreasing its revenue and diluting its value proposition. This means that based on consumers’ current accepted norms, the B2C (Business to Consumer) market for MNOs and Telcos will not get any better with 5G.

In 3G and 4G and specifically B2C, Telcos and MNOs built the networks and ran them, carrying all the CAPEX and OPEX associated with them, while everyone else enjoyed the profits. 5G, however, will open up a new B2B (Business to Business) market in Industry 4.0 players and a real opportunity for business service dollars. MNOs and Telcos need to make sure that they don’t put themselves in the situation they are in today with other members of the ecosystem and don’t invest in massive 5G buildup CAPEX while marginalizing themselves and shrinking their profit pools again.

A Painful Journey

Telcos and MNOs need to upgrade their equipment and move away from old proprietary systems provided by a handful of players to open systems based on standardized hardware and software provided by many players (existing and new). Furthermore, the new networks will be architected for a distributed world, moving away from the traditional centralized models. This is not how they had operated in previous “G” s and not what they are ready for; hence partnering with the right companies is the key to their success. It will be a painful journey initially, but it will pay dividends in the long term, provided they don’t repeat their past mistakes in new ways.

Along the way, new players will also join in and change the 5G ecosystem’s landscape. Here are some observations for your consideration:

1- Hyperscale Cloud Providers: MNOs will have to “cloudify” their networks. This is not a matter of if, but rather when. The savings are in order of magnitude in both CAPEX and OPEX when MNOs replace physical servers with cloud-based servers and move their network hardware, applications, and even parts of their operations into the Cloud. It reduces Telco’s and MNO’s Total Cost of Ownership (TCO), improves the time-to-market for rolling out new services, increases their business agility, and brings faster innovation and flexibility to enable services on-demand.

This is a massive opportunity for Hyperscale Cloud Providers, and those who are ready with network savvy professionals to help with MNOs’ migration to the Cloud will reap the benefit. In other words, cloud wars will intensify, but many service dollars will be made here. Along the way, the term “Carrier-of-Carriers” could find a new meaning in the telecom industry.

The real strategic question for the Telcos and MNOs is to handover their Cloud to the Hyperscale Cloud Providers or let them in as a part of their multi-cloud strategy and keep control.

With the decentralization of the Cloud and data centers and the rise of the Edge in 5G, additional considerations will come into play for the MNOs and Telcos. The Cloud business model of pay-as-you-go and “Everything as a Service – EaaS” will stay the same; however, Cloud “location” will change. There will still be public and private Clouds, on-prem, near prem, and centralized; however, 5G will accelerate the rise of the Edge. Edge means a distributed model where agility and customization are critical. Hyperscale Cloud Providers have typically been good at providing services at scale, in limited “centers.” Edge will be a different ball game altogether. Edge computing will provide near real-time insights and analysis by proximity to the thing, people, processes, and action generating the data and enable localized decisions and actions.

Edge computing solutions will come in many forms and have a different architecture depending on the vertical market addressed (Please note the network slicing graphics above). Every player needs to make bets on market timing, develop a vertical market strategy, and fund and execute accordingly.

At the same time, MNOs and Telcos who have been captive to traditional equipment providers and their proprietary systems have a chance to remove the monopoly based on 5G’s open standards. While diversifying their supplier-base based on open standards will be a game-changer for them, they don’t want to swap one captivity for another. They need to be careful and not become captive to a single Hyperscale Cloud Providers. This will be a strategic mistake, as most Hyperscale Could Providers also want to become B2B business service providers, and the OTT model is repeating all over again. Telcos and MNOs need to learn to become Cloud operators themselves and let Hyperscale Cloud Providers participate in their multi-cloud strategy, or they will repeat their past mistakes in a new way.

At the same time, Hyperscale Cloud Provider will find some MNOs and Telcos (Traditional and new ones) who will hand over their infrastructure keys to one Cloud player and bank on offering different B2B services while minimizing their CAPEX and OPEX. The Hyperscalers need to get ready to support this group of MNOs and Telcos with white-glove-service to entice more of these customers to join them.

2- Networking Equipment Players: 

Let’s break this group into two sets:

·        Large Networking Equipment Players: 

This group of companies includes Ericsson, Huawei, Nokia, Samsung, and ZTE. Open standards and moving away from traditional proprietary systems doesn’t necessarily mean that this group of players will lose their position in the market. Telecom business is very complex, and MNOs and Telcos are used to buying fully optimized end to end solutions that simply "work." As a friend of mine in a strategy role for one of the Mobile Operators in the US told me, “we have been buying buses off the lot for a long time, and using them to deliver customers from point A to point B. The expectation can’t be that we will now be buying tires, chassis, engines, drive trains, etc., and build the buses ourselves, no matter how many open standard interfaces were used to build those subsystems”. With the transition to 5G, MNOs and Telcos cannot compromise their operational efficiencies and customer experiences. They are used to working with their trusted partners in managing large-scale networking, IT, and Cloud operations efficiently.

A few System Integrators (SIs) would take advantage of the open interfaces and standards to integrate products from smaller players and offer an integrated system solution. However, replicating the expertise and breadth of a company like Ericsson is a very tall task. In reality, these market leaders will enjoy a prominent role for a long time to come. Let’s use Ericsson as an example of what smaller players are dealing with to put this “chasm” in perspective:

From Ericsson’s announcements: “Ericsson has been working with major MNOs and Telcos since the early days of 5G, where their first public 5G partnership announcement came in 2014. Ericsson has deployed 5G New Radio (NR) in high-, mid-, and low-bands in different urban, suburban and rural environments to support enhanced mobile broadband and fixed wireless access business cases. Ericsson’s offerings include Radio Access Network (RAN) and Core network deployments, enabled by products and solutions from the Ericsson Radio System and Ericsson Cloud Core network portfolios. Their deployments include 5G Non-Standalone, 5G Standalone, and Ericsson Spectrum Sharing technology. They also include cloud-native capabilities with 5G Core. Ericsson has also worked with service providers, universities, technology institutes, and industry partners to develop and pursue 5G business and consumer use cases. These use cases include factory automation, smart offices, remote surgery, and other enterprise and Industry 4.0 applications. As a result, they have won 100+ 5G contracts with MNOs and Telcos globally.“

This group of players needs to:

1-     Traditional Play: Continue to do what they do best, with full soup to nuts solution expertise and offerings for 5G while supporting open standards.

2-     New Front: Since 5G opens up a B2B market to MNOs and Telcos (as detailed above) for real enterprise dollars through Industry 4.0 related businesses, they should become IIoT solution providers. So far, most IoT/IIoT projects have been failing. A big part of that has been due to the diversity of the markets, players, connectivity solutions, and soup to nuts service provision, most of which plays to these players’ strength. Since 2003, I have been a supplier to the “very thin and wide” Industrial and subsequently IIoT market globally in various roles. For the IIoT market to take off at scale, it needs players such as large networking equipment players to enter and support this market, the same way they have been supporting the Telecomm industry for the last 30+ years.

5G not only solves some of the issues currently slowing down successful IIoT project rollouts, but it also enables private networks for factories and manufacturing campuses. No one is better equipped at the moment, with products and the know-how to rollout private networks like this group of companies. The breath of portfolio from these large companies can provide “a fit for purpose, high-performance, and agile private network, guaranteeing manufacturers the connectivity and security they need.”

With the XaaS and Everything-as-a-Service model becoming prevalent and some of the Hyperscale Cloud Providers getting into Edge infrastructure business, some Industry 4.0 players will ask for Telecom equipment as-a-Service, trading CAPEX with OPEX. My experience has been that large Industry 4.0 segments like Energy Production plants, Large Manufacturing Facilities, Airports, Marine, Bus, and Rail Stations, and City Municipalities are not opposed to CAPEX. Only smaller players will look for this. However, for those who ask for it, Large Networking Equipment players have the capital to support it.

·        Small Networking Equipment Players: 

These are the smaller players that collectively today have less than 10% market share. As I mentioned earlier, while based on the geopolitical issues between Washington and Beijing, Huawei and ZTE shares will be eroding in the international markets. The domestic Chinese market is large enough, however, to propel their growth. China, by itself, is a broader market than the rest of the world combined. With Huawei and ZTE vacating the western world 5G market rollout, it creates a vacuum for nimble and low-cost solution providers in the market. Most of them are smaller or non-traditional networking-player members of the O-RAN Alliance, which are awaiting technology maturity and adoption to make their moves.

For this group to seriously become a part of the overall MNOs and Telco’s future infrastructure, and even challenge the larger players, they need to work with viable SIs as a part of a much larger play, as a member of an ecosystem of solution providers. In other words, they need to gain scale through a robust ecosystem play, where partnerships matter.

This is an excellent opportunity for MNOs and Telcos to diversify away from traditional equipment manufacturers if they wish to diversify their supplier base. If not fully adopt these players, have enough of them around with a viable offering to leverage in your negotiations with the traditional large players.

3- Small Equipment (Access points, Routers, Adaptors, etc.) and Small Cell Equipment Suppliers: Based on consumers’ appetite for more data with more speed and capacity, with each “G” rolling out, the role of small cells has increased. MNOs go-to solution to safeguard the quality of service in the network’s outer Edge has been these small cells. 5G allows for smaller transceivers that cover much smaller service areas than traditional cell towers. Hence, more is required, but each costs less and consumes less power. 5G small cells are a significant part of MNOs plans, and very soon, there could be as many of them as common lamp posts, except better, blended into the landscape.

Millions of small cells will be deployed, creating revenue opportunities for companies in the small-cell 5G network market and beyond. If you were a Hyperscale Cloud Provider or an existing Large or small network equipment manufacturer, would you not want to have a piece of the pie here? There are many commonalities but a lot of differences in rolling out the small cells vs. their large-cell cousins -- enough that you need specialized teams focused on it. Could this spur a lot of M&As in this space? Acquisition of Cradle by Ericsson is an excellent example of this.

4- 5G Broadband vs. Cable: 5G broadband could create an existential threat to the cable industry, as converged services could lead to unified carriers. For a long time now, municipally-regulated monopolies have made protection for the cable industry. The emergence of 5G wireless, which can level the playing field, could change all of that. To use what is happening in the US as an example, AT&T could become a direct competitor against Comcast or Charter Communications, not just for broadband internet but also cable television. We have already seen this in multiple territories through AT&T's rollout of Fiber to the curb. However, Fiber to the curb is not readily available everywhere, but 5G small cell site could easily be put in place. With all other types of services enabled by eURLLC and mMTC around the corner, 5G could open up multiple “fronts” against cable providers.

5- Industry 4.0 (IIoT) Players: Industry 4.0 integrates IIoT, analytics, robotics, AI & ML, and augmented reality. 5G will be a crucial enabler for factories of the future by providing a unified communication platform. 5G technology will enable Industry 4.0 mission and business-critical use cases requiring high performance and low latency with high-reliability communications systems. For clarity, we usually talk about the latencies associated with 4G of approximately 20 milliseconds, whereas for 5G networks, the latency will be around 1-5 milliseconds (“near real-time”). 5G is also over 100 times faster than 4G, supporting more reliable connections. Different IIoT systems will have different network requirements; Some use cases will demand mission-critical reliability with low latency (e.g., surgical robots), while other use cases may need a higher density of connected devices, all at much higher bandwidths and peak data rates. All can use the benefits of agile private networks, especially when there are campus-wide/site related fit for purpose services are needed, supporting high-performance, high-reliability, resilient connectivity, and security.

For example, 5G enables IIoT manufacturers to create digital twins (virtual replicas of physical devices or products) via continuous streams of real-time data, which allows them to have much better visibility into their operations. This enables detecting malfunctions sooner and predicts outcomes more accurately, reducing their overall cost while improving reliability.

Therefore, IIOT and Industry 4.0 players working on components and modules, machinery and industrial automation, and manufacturing players are well-positioned to benefit from the 5G rollout. They do need to make sure their product and systems become “5G-ready”. This is where vertical-market-optimized Edge equipments will make the most impact.

6- Last but not least - Cybersecurity Issues: 2019 was considered the worst year for cybercrime, and this is just the beginning. By definition, everything is "connected" in wireless networking, hence accessible by remote bad actors from various regions, including cyber espionage and attacks from sophisticated nation-states. Earlier networks' reliance on centralized hardware-based functions offered a choke point, which could enhance security. Higher-level network functions previously achieved by physical appliances are now being virtualized in software. The 5G network has moved from centralized, hardware-based switching to distributed, software-defined digital routing, all of this leading to an increased cybersecurity vulnerability.

The benefits of software-defined anything brings along the most significant vulnerability associated with that software - security. Even if one could lock down the software vulnerabilities within the network, the whole 5G network is maintained by software, so gaining control of the network's software can control the entire network. 5G software is very complicated, broad, and distributed across the whole "system." This means 5G software is the most extensive threat surface globally and much more vulnerable than the previous "Gs."

Working in the wireless networking space until the late 2000s and then switching to Industrial and automotive, I realized the dramatically different approaches to cybersecurity within the various verticals in those days (2009/2010). Unfortunately, ten years later, there are still no security standards for the industry to follow, and many IoT and IIoT devices are manufactured with no security in mind. When it comes to cybersecurity, most low-end "smart" devices can be considered "dumb."

As someone who has been selling cybersecurity solutions and services to the market for the last five years, I can tell you that there are no silver bullets. Some of my thoughts around cybersecurity, including a holistic approach to cybersecurity, the concepts around zero-trust and defense-in-depth, and managing the lifecycle of the security, have been captured here and my previous posts on LinkedIn.

Both private companies and the government need to take broad actions to reverse years of anemic investment in security. Governments must establish agile cybersecurity regulations to reflect the new realities and complexities of our technology-driven society. At the same time, private companies must be held responsible for cybersecurity attacks due to their negligence. 

This is a massive area of concern in the industry, and with that comes a lot of opportunities for companies and developers focused on cybersecurity solutions. My next blog will be focused on this topic, specifically as it relates to 5G. 

As always, your comments are welcome.