From the exhibition’s floor to the conference’s tracks the sound message at MWC 2019 was that 5G starts to become a reality; out from the labs and trials to already first launches in the US and South Korea, with other major countries launching by the end of the year. In addition, the onset release of 5G smartphones in the first half of the year are shaping more and more the 5G promise.
Amid the 5G lovefest, however, the message that transpired was that first 5G launches are still facing technical and business challenges.
5G Challenge 1 - Spectrum Remains the Priority
The first 5G challenge is technology related. Spectrum still remains the priority, with 5G requiring larger contiguous blocks of spectrum in low- (sub 1GHz), mid- (3.5GHz) and high-range (mmW) frequency bands to achieve its potential for a large blanket coverage, secured mobility even at high vehicles’ speed and, respectively, extreme throughputs in hot spot areas. Large scale mmW are not efficient due to several known reasons (e.g. propagation challenges, limited mobility, high number of cells, back hole bandwidths requirements), making high frequency spectrums the best candidate for FWA and hot spot areas as we already have seen with a couple of US first launches. Thus, sub 6GHz spectrum with carrier aggregation deployments is the front-runner solution for the majority of operators aiming for 5G blanket coverage, smooth mobility, low latency and high throughput. Therefore, for 5G to get momentum, a lot of re-farming activities need to happen, and spectrum regulatory flexibility is still required.
5G Challenge 2 - Performance of 5G Devices Unknown
On exhibition floor was the excitement for devices’ folding display such as Huawei’s Mate X and Samsung Galaxy Fold; and for the LG V50 double screen; but these devices are just “5G like” in the sense that they provide pixel resolutions suitable for the 4K video content. The first expected devices to support 5G NR (sub 6GHz and mmW) are NetGear (on AT&T network), HTC’s 5G Hub (on Sprint, Telstra network), Motorola 5G Moto Mod (Vzw network), Samsung Galaxy S10+5G, and Huawei’s OnePlus 5G. So far, only scanner-based testing solutions and crude prototype devices have been used to evaluate coverage, capacity and throughput performance and gains vs. counterpart 4G/LTE as expected from MU mMIMO (sub 6Ghz) and beamforming (mmW) systems. However, much still remains to be seen:
- how fast these new 5G commercial devices are going to detect and acquire beams
- how smooth beam switching is achieved in handover scenarios
- how well multi-module device based antennas perform in challenging scenarios (e.g. mmW, especially indoor)
- how the LTE-NR synchronization happens at device level in dual connectivity scenarios, and
- what is the ultimate QoS/QoE performance (e.g. throughput, latency, QoE video quality)
Huawei’s talk nicely quantified 5G devices’ challenges versus their 4G/LTE counterparts in the following categories: 5x more computing power, 2.5x power consumption, 1.3x complexity due to multilayer, multimode/multi connectivity, and multiband, especially the devices’ front-end filtering which needs to support simultaneously low and high frequency (LTE-NR mmW) dual connectivity scenarios, while coping with heat dissipation (mmW). The showcased simulated demo at Qualcomm’s booth put in positive perspective some of these aspects. Snapdragon X50 in the indoor 5G mmW 28GHz mobility using 3 module device based antenna forming 16 beams was showed to usefully mitigate the effects of hand-blocking as well as to better exploit reflections to extend mmW coverage for a uniform coverage and robustness to overcome blockage and orientation change. In addition, Qualcomm simulated demo showed that outdoor 5G NR mmW deployments can also be used to provide select indoor connectivity.
However, we still have to wait and see the 5G commercial devices in the field, with real-time data collection and logging that will allow in-depth and user centric performance analysis.
5G Challenge 3 - AI and ML Needed for 5G Networks Performance Management
So, the third 5G challenge that emerged is the cost efficiency of 5G networks which are distinct from previous generations because of the level of heterogeneity, flexibility and automation inherent in their design. The cost dynamics of 5G networks will therefore not only be influenced by traditional factors (e.g. capacity and coverage), but also by new factors such as network flexibility and network ownership. Some of these factors are already being addressed in 4G networks (for example, NFV/SDN for network flexibility and edge computing for low latency capabilities), but their impact on the cost of 5G network rollout and operations is not clearly defined yet. The noticeable trends toward answering to these concerns is the introduction of new models of network ownership (e.g. private 5G networks), new ways of building networks (e.g. using open source concepts and cloud- based solutions) and new network performance management approaches (e.g. using AI-based automation).
For example, part of the new way of building networks are the two RAN options still under discussion: vRAN and O-RAN; both with pros and cons., but better than multi-vendor solutions. The final choice is not here yet, and various vendors and operators are biased towards one or the other, or a combination of these, but major RAN vendors are still expected to emphasize the cost benefits of site simplification to counter vRAN or O-RAN appeal.
When it comes to new network performance management approaches the trend is obvious and played already at various levels of magnitude by vendors and operators alike; and this is AI, ML and consequently “intelligent connectivity”, this year’s MWC motto. All of these involve the existence of intelligent, AI-based devices and autonomous agile networks which adapt and respond to the traffic on the fly, improving the customer experience through greater learning of customer behavior. However, the AI-based network performance management is not limited to 5G, but needs to cover 4G and the other legacy networks. Intelligent automation for OPEX savings is crucial for freeing up resource towards 5G deployments, which also have to rely on AI based CAPEX strategies. From the exhibition floor to the conference tracks surfaced the vendors’ (e.g. Ericsson, Nokia), 3rd parties’ (e.g. Amdocs) as well as operators’ (e.g. Telefonica, Elisa) efforts to develop AI-based solutions for intelligent network performance management with use cases covering data driven capacity planning, operational efficiency, dynamic customer experience and network performance prediction and optimization, and customer care.
AI, ML and, consequently, “intelligent connectivity” were also the topics laying at the core of the fourth 5G challenge emerging from MWC 2019 discussions.
5G Challenge 4 - Network Slicing Provisioning and Quality Assurance Not Ready for IoT Verticals
The fourth 5G challenge has been intensively discussed during MWC conference’s tracks and extensively displayed on the exhibition floor and emerged from the need for 5G “big bang” which will maximize data revenues and is expected to enable “real” 5G.
So far, the single available use case for 5G is eMBB, and just until lately “faster and better” didn’t make the best case for 5G toward the consumers. However, the low latency and high bandwidth enable the technical feasibility of AR/VR applications and, consequently, the cloud gaming with Netflix- like subscriptions which started to show a significant business potential. Therefore, MWC 2019 closed on an optimist tone regarding 5G eMBB for 8K video and cloud gaming both in FWA and mobile scenarios.
However, these are neither enough for 5G to get momentum nor to maximize operators’ data revenue. In addition, with all the technical challenges that we mentioned to be still in place it is very clear that there is a long way to go until 5G hits the critical mass; GSMA Intelligence report published after MWC 2019 shows that by 2025 59% of connections will be still 4G while only 15% are expected to be 5G.
Therefore, it is clear why the MWC 2019 event was crowded with use cases and applications showing operators’ move beyond their traditional telco businesses (mobile and fixed) and partner with various verticals towards new revenue streams. While this strategic play has different approaches, timelines and scales, the predominant drivers are the rise of IoT, the evolution of the content ecosystem, and the onset of a new era of connected devices.
Industrial IoT, automotive and healthcare were the most predominant ones. And just listing very few examples: various demoes of manufacturing robots powered by 5G mmW NR; self-driven connected car powered by proprietary AI technology; demoes of health care use cases, such as remote, real-time diagnosis powered by 5G sub 6GHz network. And maybe the most exciting of all was the remotely mentored surgery showcased in real time, with the doctor in the MWC conference room and the patient and assisting doctor and nurses in a Barcelona downtown hospital. The demo used 5G sub 6GHz NR connectivity which enabled the low latency and video bandwidths for the doctor to direct in real time the surgery.
All very exciting, but still ahead. 3GPP Release 16 aimed to unleash URLLC and mMTC use cases is postponed to mid 2020. In addition, network slicing dynamic allocation, provisioning and quality assurance to ensure the optimal network core and RAN resources with the specific QoS requirements for these kinds of verticals is not available yet; neither are the test tools to validate minimum QoS requirements and to verify SLAs violations. However, on the exhibition’s floor, a couple of demoes from vendors (e.g. Nokia, Ericsson) showed their efforts towards AI-based network slicing orchestration scenarios.
In a nutshell, the lessons we learnt at MWC 2019 cover the following main topics. The hurdles of 5G NR deployments from spectrum availability in different frequency bands, to coverage planning for in demand traffic areas and sites verification to the need of commercial 5G devices’ field performance evaluation. Then there is the exponentially growing need for AI and ML based solutions for legacy and 5G networks performance’s management, optimization, and ultimately prediction. The emerging IoT verticals rise the need for intelligent mobile edge test solution as well as for solutions for network slicing provisioning and quality assurance.
Learn more about our 5G strategy and how our experts are helping launch the first deployments around the world.
