Just last month, I had the pleasure to speak at the 4G/5G Summit in Hong Kong, one of the largest wireless ecosystem events in the world, hosted by Qualcomm Technologies. It was attended by more than 2,300 representatives from all parts of the global industry — operators, OEMs, infrastructure vendors, device manufacturers, service providers, and more, further attesting to Qualcomm’s leadership and commitment to working with the entire ecosystem to make 5G a success.
As we approach 2019 — the year 5G NR comes to smartphones, one of the questions on the top of everyone’s mind is “how can we deploy 5G NR in a timely and efficient manner?” At 4G/5G Summit, I discussed our Qualcomm Technologies Engineering Services Group (ESG), which works with operators around the world to address deployment challenges with new wireless technologies. I thought the following key takeaways from my presentation could help answer this pressing question.
Co-siting with outdoor LTE infrastructure can provide significant coverage at 3.5 GHz
The most practical approach to deploy 5G NR, whether in sub-6 GHz or mmWave bands, is to leverage as much of the existing infrastructure as possible. Not only can co-siting accelerate the deployment process (e.g., avoid conducting extensive site surveys and applying for new site licenses), it can also reduce capital investments (e.g., reuse/upgrade fibers vs. deploying new ones). And with the recent LTE densification to support Gigabit LTE, driven by LAA small cells, it makes this strategy effective and cost efficient.
For more clarity, we cooperated with global operators and looked at their existing LTE outdoor network deployments and simulated coverage by co-siting 5G NR operating in 3.5 GHz with LTE. The results were encouraging, as all three example cities shown (Figure 1) — with different site densities — showed comparable outdoor coverage with LTE. While the outdoor-to-indoor coverage is relative lower, it’s still significant and can be further improved with more indoor small cells. And let’s not forget that even with LTE operating in lower bands (e.g., 1.9 GHz), the out-to-in coverage also requires indoor solutions in many cases. Additionally, the coverage is driven by a certain expectation on cell-edge throughput delivered (e.g., 30 Mbps over 100 MHz assuming minimum spectral efficiency of 0.3 bps/Hz), which is still significantly higher than LTE today, and an operator could achieve more coverage by adjusting for the target performance they want to deliver to their customers.
To estimate coverage with 3.5 GHz accurately, another important consideration is that given this is a new band, it is recommended to develop tuned models using actual measurements in representative morphologies (e.g., dense urban with more high-rises vs. rural areas with lower height buildings) and clusters. Additionally, depending on deployment scenario, some coexistence aspects such as coordination of TDD slot structure, impact of near-far effect also need to be considered.
Similar LTE co-siting strategy also delivers significant outdoor mmWave coverage and performance
Perhaps the more interesting question is how 5G NR mmWave will perform in outdoor settings and what coverage can be achieved via co-siting. This is a more challenging analysis as we need to also take into consideration that mmWave is more susceptible to blockage, and a meaningful study requires the use of accurate 3D models of cities that include information such as terrain, buildings, and even foliage.
Again, we collaborated with global operators and conducted comprehensive simulations for several key metropolitan cities across the globe (Figure 2), looking at different LTE deployment densities, city layouts, and mmWave frequencies. The findings are summarized in the figure below. We saw significant mmWave outdoor coverage for both downlink and uplink just by co-siting with existing LTE infrastructure, and as expected with mmWave, the median burst rate is well over 1 Gbps with many cities delivering beyond 2 Gbps, capable of fueling new and enhanced mobile experiences. Even at the cell edge with suboptimal signal quality, mmWave can deliver throughputs of over 300 Mbps, which is remarkable compared to just a few Mbps that today’s networks are capable of at the cell edge. To achieve even better throughput and coverage, operators can further densify their networks. In fact, many operators have been deploying small cells to leverage capacity and performance benefits from technologies like LAA (licensed assisted access), and this dense heterogeneous deployment allows delivering near-ubiquitous outdoor coverage.
Deployment considerations for active antenna systems in 3.5 GHz
Massive MIMO is a game changer for operating 5G NR in the higher mid-bands such as 3.5 GHz. We are already seeing massive MIMO being used in 4G networks today, and its design is further enhanced in 5G NR with features such as TDD reciprocity (i.e., utilizing UL-SRS, CSI-RS). With it, operators can tap into more bandwidth and capacity, and it also enhances coverage at the cell edge to deliver a more uniform mobile experience as users move between sites.
The performance achieved with an active antenna system is dependent on the number of supported sync and reference signals, user distribution in the cell and its morphology, and the infrastructure vendor-specific implementations. As the number of antenna elements is increased, beamforming performance improves with the tradeoff of higher weight and cost.
We conducted simulations to compare capacity gains utilizing different antenna configurations such as number of elements and antenna shapes. For dense urban deployments, we found that horizontal and vertical traffic distributions can benefit from square shaped antennas; thus, we recommend a 128×2 massive MIMO antenna with at least 32 transmit and 32 receive antenna ports, with the configuration of 8 (horizontal) x 16 (vertical) x 2(x-polarization).
Most of the early 5G deployments in 2019 are expected to utilize the non-standalone (NSA) architecture, as it allows operators to leverage their existing LTE network with ubiquitous coverage, and opportunistically serve data on the 5G NR network via dual-connectivity. This not only minimizes the time to deploy 5G NR (as only a minimal software upgrade is required for the LTE core network and base stations) but also provides a more seamless user experience as users move in and out of 5G coverage.
For voice services, Voice over 5G NR (VoNR) is already in the Release 15 3GPP specifications for Standalone (SA) deployments, and we expect further optimizations to come in future 5G NR releases. As we witnessed in the VoLTE adoption cycle, it took the industry a few years to plan, test, and commercialize new voice services, and we believe VoNR will follow a similar timeline. And since LTE is the foundation for 5G NR deployments, delivering many essential services from Day 1, we expect VoLTE to remain the 5G voice solution for many years to come.
Making the 5G vision a commercial reality is no small feat. It requires many years of R&D and working closely with wireless ecosystem leaders on standardization, and then on product development with extensive interoperability testing to address the many challenges and complexities of 5G. The significant progress made in 5G commercialization is evidenced by the many recent milestones such as the first 5G NR sub-6 GHz and mmWave OTA calls using a mobile formfactor device. This was in addition to the announcement of our smallest-to-date Qualcomm QTM052 5G NR mmWave antenna module, which is 25% smaller than the previous version we announced in July of this year, proving that 5G NR mmWave is viable for the smartphone formfactor. But we’re certain that when 5G becomes a commercial reality starting in 2019, all of the ecosystem’s efforts will have been worth it.