Augmented reality over 5G
There is undoubtedly a lot of anticipation in the mobile community around the uptake of AR services over cellular networks. At this early stage in the emergence of a market that is projected to become important for service providers, we explore how the ecosystem is developing.
- AR represents a major branch of the XR taxonomy that will place significant demands on 5G mobile network dimensioning.
- As the AR ecosystem develops, traffic arising from AR usage could significantly exceed the current mobile traffic forecast.
- When AR reaches mass-market adoption, service providers will need to take a stepwise approach to handle the traffic growth, including investments in Radio Access Networks (RAN), time-critical features and service-enabling platforms.
XR encompasses a taxonomy of technologies, including VR and AR
The VR branch is comprised of a range of immersive techniques which share several attributes. These include closed headsets and cameras directed inward to capture and model their subjects. Due to its closed nature, VR needs a confined area for its subjects to experience immersive services comfortably while using a closed head-mounted display. Other related technologies include 3D conferencing and volumetric video.
AR is applicable both indoors and outdoors for a large range of services for consumer, enterprise and industrial use. AR requires an open display with a view of reality which can be augmented with relevant data, making outdoor applications particularly relevant, yet also reliant on sufficient mobile coverage, capacity and performance.
The AR ecosystem
To meet market expectations, all parts of the AR ecosystem – including devices, apps, networks and edge compute – need to evolve with new capabilities, increased performance and greater efficiencies. While the devices receive most of the attention, the availability of enough suitable spectrum and investments needed to build out the networks are also significant.
AR devices
The development of head-mounted displays, in the form of glasses, is highly visible on the road to emerging AR applications and services. Behind the form factor lie questions as to how critical functions will be split and interoperability ensured between the headset, eventual tethering platform (smartphone), network edge and the cloud. This includes object detection and tracking, simultaneous localization and mapping (SLAM), and rendering of video streams combining reality and data augmentation.
The technology required to create smart glasses for mass-market adoption has progressed considerably in the last few years. Several prototypes, as well as commercial solutions, have emerged. As of 2022, however, AR headsets have not yet gained mass-market success. The technology components of the devices must mature enough to allow for a practical and attractive consumer product. Useful applications must also be developed and integrated over the ecosystem consisting of wearable devices/glasses, connectivity devices, and ultimately cloud-based compute resources. Additionally, high power consumption and low battery capacity are current challenges.
Figure 18: AR over 5G
The technology is not yet mature enough to enable comfortable standalone devices. Key factors for mass-market adoption include device form factor and connectivity. Mobility requirements drive cellular connectivity, but this is limited by form factor and power, which necessitates a paired connectivity device, such as a smartphone. Smartphones will continue to be connectivity hubs for personal devices in the near term.
The next wave of AR glasses will likely be more attractive, but will continue to rely on a connectivity device. Higher chip density and lower power consumption will be essential for size and weight considerations in glasses, as well as heat dissipation. Improved coverage and latency in 5G networks will allow more compute offload, and AR devices will have further reduced power consumption and better battery capacity, thereby enabling improved form factors. In the meantime, offloading processing to tethered smartphones will continue.
Application handling
Consumer categories with high potential for wide uptake in the intermediate term include gaming, entertainment, and retail. AR technology will also be used to improve enterprise and industrial operations in many areas, such as private networks, virtualized work, industrial automation, design, maintenance and public safety.
There is considerable work behind the scenes on the network architecture necessary for handling the requirements for delivering the emerging applications. Each application will place quality-of-service (QoS) demands on end-to-end service delivery over the networks connecting devices. These demands include uplink and downlink throughput, round-trip latency and reliability. Networks will handle the demands by setting up QoS flows that segregate traffic into categories ranging from best-efforts to performance guarantees.
Traffic from services requiring time-critical treatment will be segregated from eMBB traffic and assigned a different QoS flow. Ultimately, application developers will use tags provided by either device ecosystem players through their operating systems or service providers. The traffic categories will then be mapped to 3GPP QoS flows to enable interoperability across networks.
Service providers will be able to differentiate their subscription plans to map premium users’ traffic to packet treatment with low latency features while mapping standard users’ traffic to the default packet treatment.
5G network traffic
Our forecast for traffic growth includes an assumption that an uptake of XR-type services will happen in the latter part of the forecast period. The current forecast (see page 22) includes estimates of moderate uptake in the forecast period. This includes a limited effect on total traffic so far, with the most significant impact on uplink traffic ratio. To understand the forecast, it is important to be aware of the following factors:
- The forecast relies on a foundation of measurements in live networks worldwide. These measurements include indications of traffic types, including – but not limited to – voice, messaging, streaming media, browsing, social media and e-commerce.
- The global traffic forecast depicts the total mobile traffic in EB as well as the average mobile data traffic per device in GB across all regions and users over the course of a month. Actual traffic in a country, city or area and across urban, suburban and rural landscapes shows very large variations from the global averages.
As the AR ecosystem develops, traffic arising from AR usage could significantly impact the current forecast. The amount of traffic that will be generated over mobile networks, in addition to mobile broadband and fixed wireless traffic, will depend not only on the uptake and utilization rates of the applications, but also where the critical functions mentioned in the “AR devices” section take place. In the next wave, it is expected that compute functions including rendering of the composite video will take place either on a tethered smartphone or the AR headset.
As the augmentation objects become larger and rendering becomes more demanding, the work will increasingly need to be offloaded to network edge and cloud compute resources.
This is expected to put growing demands on networks especially uplink throughput and latency. The ability to meet those resource demands depends on spectrum employed, as well as network dimensioning.
As more demanding use cases start to gain significant uptake and increasing amounts of compute function is offloaded to the network edge and cloud, service providers will need to respond by deploying features and solutions to handle time-critical communications.