When the mobile industry is disrupted, it typically responds by inventing a new network technology. Sometimes this pays off, but often it doesn’t. Today the mobile industry is confronted with stagnant revenues and disruptive service demands. In response, the industry has invented 5G to deliver enhanced mobile broadband, ultra-low latency, mission critical connectivity, and “network slicing” capabilities to segment resources according to service demands. To succeed, 5G must create sufficient value for mobile operators. This value will not come merely from 5G performance, but will depend on much-needed operator transformation to capitalize on changing mobile service demands.
5G can be a catalyst for value creation - but only if mobile operators let it
Today much of the value from mobile services is captured by web platform providers, such as Facebook, Google and Uber. These players have pioneered cloud architectures with virtualized infrastructure and agile operations and business models that capitalize on the growing ubiquity of smartphones, and they curate rather than control critical market assets. For example, Facebook and Google are essentially media companies, even though they don’t own any media. Uber is a transportation company even though it does not own any vehicles. This contrasts the vertically integrated strategies that are typically pursued by mobile operators. However, if operators play their cards right, 5G can be a catalyst for change to propel operators towards platform-centric business models.
Ostensibly many of the core 5G capabilities enable operators to do what they do better, instead of providing a catalyst for transformation. Enhanced mobile broadband (EMB) enables faster networks. Ultra-low latency and ultra-reliable connectivity increases the variety of services that operators can support. If operators merely focus on these capabilities for 5G, it will fail. Nevertheless, 5G also capitalizes on end-to-end virtualization with network slicing for allocating network resources according to service demands. If implemented right, network slicing could provide operators with the necessary fabric to deploy network-centric service platforms, with the aim of curating third party services and applications. Today 5G standards address only part of what is needed for network slicing. Other requirements include:
- Advances in business and operational support systems to enable end-to-end self-service capabilities, which depend on modular and autonomous ecosystems, with sophisticated management and orchestration capabilities, and;
- Changes to organizational structures, to eliminate operational silos and incentivize the adoption of autonomous operational techniques.
In general these requirements are more challenging to implement than 5G network slicing functionality and given the tremendous inertia of status quo, are best initially deployed as overlaid or shadow environments to existing ecosystems.
EMB network constraints might be a blessing in disguise
Global 5G hype is following a well-trodden path that focuses on the theoretical peak data rates achievable with 5G/EMB. Not surprisingly, these peak rates are vastly exaggerated and come with many caveats. To enable the peak data rates for EMB, 5G aggregates large tracts of radio spectrum in either sub-6GHz bands or at much higher frequencies between 10 and 100 GHz (i.e. millimeter and centimeter-wave). When millimeter and centimeter wave technologies are used, coverage is typically limited to small-cell ranges, and mobility to pedestrian speeds. 5G-EMB solutions that operate in sub-6GHz bands rely on an ability to aggregate massive tracts of spectrum and have challenging economics in capacity constrained environments. Furthermore, advancements towards 4G, with what the industry refers to as 4.5 and 4.9G, will achieve peak rates comparable to 5G in the same sub-6GHz spectrum.
While centimeter and millimeter wave technologies are severally coverage constrained relative to traditional mobile services, they have the potential to catalyze industry transformation by advancing autonomous network management, and simplifying network architectures and their deployment. In particular, 5G/EMB requires sophisticated real-time radio management and optimization capabilities that are predicated on advanced self-organizing-network (SON) principles. Network architectures can be simplified by collapsing access and back-haul domains. Furthermore, since 5G/EMB is essentially a small-cell, limited mobility wireless technology, it can support self-install distribution models to curate network coverage and innovative services and applications in localized environments, such as public venues and enterprise campuses.
Ultra low-latency forges an uncertain path towards future services
Ultra-low latency connectivity is not required for today’s mobile services, but is needed for emerging services such as autonomous vehicle connectivity and the tactile internet. The tactile internet spans a broad range of emerging services with complex value chains and includes compelling (albeit futuristic) applications, such as augmented reality, and others that are more ominous like remote surgery.
However the role of 5G for autonomous vehicles and the tactile internet are far from certain. Tactile internet applications will take many years to achieve mass market adoption and might be implemented primarily with local area connectivity solutions and not require 5G. Autonomous vehicles are fast becoming a reality and depends on ultra-low latency connectivity for coordinating vehicles. However this connectivity is typically peer-to-peer as opposed to network based and 5G is arguably over-designed for the job. The automotive industry has already developed dedicated short range communications (DSRC) for low latency V2V and V2X connectivity. DSRC has the potential to become the de-facto standard for autonomous vehicles, rather than 5G.
To succeed against competitive ultra-low-latency solutions, we believe that the value proposition for 5G must be more than just connectivity. 5G efforts must be closely aligned with mobile edge computing initiatives, and mobile edge computing architectures must natively incorporate end point devices as part of the “active edge”.
5G brings IoT connectivity on steroids - but is it really needed?
The Internet-of-Things (IoT) is expected to proliferate in coming years and herald a massive growth in the number of connected devices. Traditional cellular technologies that have been designed for mobile users are constrained in the number of devices that can be supported by each base station and incorporate unwieldy core network functionality that is not required for machine type communications. However, low power wireless access (LPWA) network technologies and 3GPP standards like LTE-M1 and M2 have been designed to support massive connection densities, with significant improvements in device energy efficiencies.
The 5G standards intend to support over 1 million devices per base station, which corresponds to between 20 and 40 thousand simultaneously connected devices per square mile. While we expect the number of connected devices to increase dramatically over the coming years, we do not believe that 5G is necessarily required. LPWA and LTE-M1 and M2 technologies are currently more than adequate and will be improved as device densities increase. Rather than focusing on connection densities, 5G might be better positioned in the market with other capabilities, such as network slicing.
5G techno-lust is compromising value creation
It is easy to get wrapped up in the technical sophistication of 5G and lose sight of whether or not it delivers sufficient value for mobile operators. We do not believe that this value will come from the “speeds and feeds” that 5G enables. Rather it will come from the role that 5G can play in catalyzing mobile operator transformation, with architectures, operations and business models that parallel those pioneered by web platform companies like Facebook, Google and Uber. In addition, 5G cannot be developed in isolation of other transformation initiatives associated with cloud and network virtualization, edge computing, and business and operational support systems.
Phil Marshall is the chief research officer of Tolaga Research