Test Considerations for 5G New Radio

Test Considerations for 5G New Radio WHITE PAPER With the implementation of 5G over the next decade, both network equipment manufacturers (NEMs) and device manufactures will face new challenges in testing their hardware, software, and end-to-end deployments. 5G technology, as currently envisioned, is quite different from 4G/Long Term Evolution (LTE) and will bring together some of the most challenging aspects of existing approaches. It will also introduce new challenges, described further here. As these challenges are overcome, the technology will usher in a host of potentially game-changing applications, depicted in blue in Figure 1. DELAY 1ms Autonomous Driving Augmented Reality Tactile Internet Virtual Reality Services that can be delivered by legacy networks 10ms Disaster Alert Real-Time Gaming Multi-Person Video Call Services that could be enabled by 5G 100ms Automotive ecall Device Remote Controlling Bi-Directional Remote Controlling First Responder Connectivity Fixed Nomadic On-The-Go 1,000ms Monitoring Sensor Networks Personal Video Cloud Streaming Wireless Cloud Based Office <1 Mbps 1 Mbps 10 Mbps 100 Mbp >1Gbps Figure 1. Potential game-changing applications enabled by coming 5G standards. Machine-to-machine connectivity BANDWIDTH THROUGHPUT Page 1

Keysight is leading the way in delivering first-to-market 5G solutions. The hardware and software solutions pave the path from 4G/LTE test architecture to true next generation 5G test architectures that will help both operators and their equipment suppliers validate their designs and network configurations to ensure devices and networks perform as expected. Use Cases for 5G Test architecture must evolve and adapt for 5G, effectively handling the three major use cases identified by the Third Generation Partnership Project (3GPP). The 3GPP is responsible for the technology evolution and works with the better-known GSM Association (GSMA), which represents the operators. The three use cases are: 1. Enhanced Mobile Broadband (embb). This is an evolution of today s 4G use case streaming video, conferencing, and basic broadband connectivity but with greater emphasis on uplink capacity. The downlink speeds will increase by an order of magnitude 10-20 Gbps per cell. 5G is also expected to replace fixed broadband connections in some cases. In fact, this use case will be the very first 5G deployment. 2. Massive Machine Type Communications (mmtc). This is an evolution of today s Internet of Things (IoT), but with orders of magnitude more endpoints. Solutions here focus on power efficiency and the ability to support extreme node density. 3. Ultra-Reliable and Low Latency Communications (URLLC). This is an entirely new set of applications made possible by low latency, high reliability, and in many cases, high bandwidth. They include virtual and augmented reality, remote real-time surgery (tactile internet), and autonomous vehicles. Connections must be secure and maintained at high speed. Extreme capacity Extreme data rates embb Area Traffic Capacity (Mbit/s/m 2 ) Peak Data Rate 10 1 20 High Importance Medium 10 1 Low 100 User Experienced Data Rate (Mbit/s) 1x 3x Spectrum Efficiency Network Energy Efficiency 100x 10x 10 5 10 350 500 Mobility (Km/s) Ultra-low energy Ultra-low complexity Ultra-high density Deep coverage mmtc 10 6 Connection Density (device/km 2 ) 1 Latency (ms) URLLC Strong security Ultra-high reliability Ultra-low latency Extreme mobility International Mobile Telecommunications system requirements for the year 2020 (IMT-2020) mapped to 5G use cases. Page 2

Luckily, not all of these use cases will appear at once, enabling test architectures to evolve over the 5G adoption time frame. 3GPP Roadmap Looking at requirements, as described by the GSMA and adopted by the carriers, a basic tenet is the technology will support 10-20 Gbps per cell with potentially 1 Gbps real (or perceived) downstream bandwidth to a user at less than 1 ms latency. This is an important distinction, since LTE was advertised as having high peak data rates, but the end user rarely, if ever, experienced them. Even fixed broadband networks must rise to the challenge of providing these rates and, thus, the required new architectures. This implies a network that is: Capable of supporting much greater data volumes Secure Fluid and flexible to support potentially trillions of devices (i.e., IoT) Adaptable to the application versus the other way around An effective way to understand these requirements and the required network evolution is to map them to the phases of the 3GPP s International Mobile Telecommunication system requirements for the year 2020 (IMT-2020). There are two phases to this. Phase I is focused on early availability and an evolution of existing LTE deployments and is aligned to 3GPP Release 15. Release 15 initially defined a non-stand-alone mode (NSA), where 5G is anchored by a 4G LTE core network. This was standardized by 3GPP in December 2017. Standalone mode is expected in the final 3GPP Release 15 later in 2018. Phase 2, will be defined by 2019, if not sooner. Some operators are claiming first commercial deployments of 5G towards the end of 2018 and into early 2019. Are there test architectures for this technology shift? One way to look at this is through the use cases instead of the technology itself. For example, model the top 10 user scenarios, and determine the architecture required. This is important, since developing these capabilities to validate an (IoT) sensor, an autonomous vehicle, streaming of a high-definition (HD) movie, or a remote surgery are all very different. Endpoints appear and disappear at high rates, cell-site complexity grows with network sharing, and even the radios deployed and the bandwidth required for the visibility traffic itself will require new ways of thinking. Page 3

Implications of 5G The GSMA has listed a set of goals for 5G, all of which will not be attainable at once due to limited technology or commercial limitations. As mentioned above, aggressive goals of 10 Gbps data rate and 1 ms latency will be a challenge, but this is only the beginning. Aspirational goals include a perception of 99.999% availability and 100% coverage, as well as a 10x reduction in network energy usage and a 10-year battery life for low-power devices. All this at 1000x the effective bandwidth per unit area or, in other words, base stations that can support 1000x the capacity of current base stations. This comes at a cost of requiring much higher frequencies that can support higher bandwidth channels. These higher frequencies imply much higher density cell deployment (i.e., microcells), which then require a much higher density back haul infrastructure. The user perception of availability is also impacted by roaming, made more difficult by the above network requirements. However, looking at the three use cases, roaming may really only be applicable to enhanced mobile broadband where users may also fall back to 4G/LTE. One of the most difficult challenges of 5G deployments will be latency. Both bandwidth and coverage have known engineering solutions, including the deployment of additional hardware or the allocation of additional frequencies. But near instant, fail-safe latency will be challenging for use cases such as virtual reality (VR) and distributed control, require real-time response. Currently, LTE latency may be in the range of 40 ms or more, a noticeable impact in application performance and user perception. Coupled with PerfectStorm, XAir2 helps operators and NEMs prepare for 5G by supporting many facets of pre-5g support for LTE in unlicensed spectrum (LTE-U)/Licensed Assisted Access (LAA), Narrow Band IoT (NB-IoT), multiple carrier aggregation on downlink and uplink, 4x4 MIMO, 256 QAM, dual connectivity, and more. For 5G, reducing latency by an order of magnitude implies that some of the servers and intelligence will need to be within proximity of the cell site, and the driving factor will be the cost of delivering the service versus the net gain. Some of the work involving edge data centers and fiber buildouts in support of the distributed cloud by the cellular providers align to these requirements. There may be different approaches to the edge and near edge, depending on who is building the network and providing the content. Page 4

A number of other considerations come into play with 5G deployments, issues that were either not applicable to 4G/LTE or that were on the periphery. For radios, 5G will extend operation into much higher mmwave frequencies with current investigations up to 70 GHz as shown in figure 3. Technologies, including beam forming and higher-order multiple input, multiple output (MIMO), come into play, and any deployment will need to avoid interference with existing technologies, such as Wi-Fi and other wireless communications systems; like satellite. It is also possible that different services, based on different over-the-air requirements, may use different frequencies. The basis of this is defined as the 5G New Radio (5G NR), with first availability part of Release 15. Adding to test requirements is a newly defined Non-Standalone 5G NR planned for 2019. Any test hardware will need to support these new frequencies and test for interference. -10 GHz 10-20 GHz 20-30 GHz 30-40 GHz 40-50 GHz 50-60 GHz 60-70 GHz 70-80 GHz 80-90 GHz Frequency ranges proposed by regional groups Europe CIS (RCC) 24.5 27.5 31.8 33.4 40.5 43.5 45.5 48.9 66 40.5 48.6 50.2 25.5 27.5 31.8 33.4 39.5 41.5 45.5 47.5 50.4 52.8 66 71 71 76 76 81 81 86 86 Keysight s design and test solutions for NEMs, operators, and device manufacturers enable characterization of RF performance, validation of protocol, and over-theair (OTA) tests to ensure products work through the ecosystem: Test designs with 5G compliant waveforms Simulate and validate Massive MIMO and beamforming patterns Test for throughput on the network Validate over-the-air performance. Arab (ASMG) No specific frequency bands submitted, opinions that above 31 GHz should be targeted. Africa (ATU) 26.5 27.5 45.5 50.2 7.057 10.5 17.3 23.6 24.25 31.8 33.4 37 43.5 50.4 52.6 55 76 81 86 The Americas (CITEL) 23.15 23.6 31.8 33 47.2 50.2 10 10.45 24.25 27.5 29.5 37 40.5 45.5 47 50.4 52.6 59.3 76 47.2 50.2 Asia-Pacific (APT) 25.25 25.5 31.8 33.4 39 47 50.4 52.6 66 76 81 86-10 GHz 10-20 GHz 20-30 GHz 30-40 GHz 40-50 GHz 50-60 GHz 60-70 GHz 70-80 GHz 80-90 GHz Agreed frequency ranges to study 24.25 27.5 31.8 33.4 37 43.5 45.5 50.2 50.4 52.6 66 76 81 86 Figure 3. Frequency ranges being studied for identification at World Radio Communication Conference in 2019. The massive increase in the number of devices will only be served by a combination of greater over-the-air bandwidth, antenna designs that efficiently handle this bandwidth and subscriber density, and addressing space availability with Internet Protocol version 6 (IPv6). More interesting and challenging is the concept of network slicing, described as dynamically and automatically expanding and contracting network capacity to support a given use case, the adaptation and coordination of network segments to deliver a certain flow to a specific endpoint at a given specific time. Page 5

In either a single operator or a shared deployment, any test architecture will need to properly characterize how the network is partitioned, the logical slices, and the services offered over each with regards to latency, throughput, and availability. Given the role software and services play in this architecture, we can use the term Software-Defined Test (SD-Test) to reflect this software-centric test architecture. Here, it must match the 5G architecture itself, an evolved software-defined networking (SDN)/network function virtualization (NFV) infrastructure enabling stateless network functions that offer total flexibility based on a micro-services architecture. Enhanced Mobile Broadband In embb, the user experience is based on the perceived data rate, both upstream and downstream. This implies sufficient capacity, which is also a function of spectrum and network energy efficiency. Testing must scale to the expected 100x network capacities and 100x effective data rates; model radio performance in some of the new frequency bands in sub-6 GHz and mmwave; model small cell access/macro cells; and simulate the different applications that will include HD video and video conferencing. embb will utilize existing and new technologies to achieve the expected extreme data throughputs: 1. Spectrum. 5G NR will extend the radio channel into mmwave frequencies and combine unlicensed spectrum to increase the possible channel bandwidths. With more channel bandwidth, more data can be sent through the channel. 2. Carrier Aggregation. LTE-A Pro supports aggregating up to 32 carriers to form larger bandwidth blocks. 5G NR initially specifies aggregation with up to 16 component carriers with an aggregated channel bandwidth up to 1 GHz, and greater channel bandwidths expected in the future. 3. Multiple Antennas. MIMO in LTE-A Pro and Massive MIMO configurations in 5G will enable greater throughput through multiple, independent streams of data. As 5G evolves it s expected to see base stations with 100 s of antennas. 4. Modulation Density. LTE-A Pro and initial 5G NR specify up to 256 quadrature amplitude modulation (QAM) to increase data throughput by encoding more bits per symbol. With 5G, this can go up to 1024 QAM and beyond. Page 6

Being first to market, Keysight and Ixia embb solutions span the ecosystem to ensure designers are testing for the use cases specified by operators. Keysight s 5G Waveform Generation and Analysis Testbed is 5G NR-ready and enables characterization of 5G NR devices and equipment from RF to mmwave frequencies with precision and modulation bandwidths up to 2 GHz. It includes multi-format testing with wide bandwidth for testing of carrier aggregation, modulation quality, and coexistence in a compact form. Keysights s 5G Protocol toolset enables a seamless RF and protocol workflow approach that enables device manufacturers to efficiently develop and test to the latest 4G and evolving 5G standards in a single solution. Ixia s IxLoad LTE XAir2 permits service providers, equipment and chipset makers, and enterprises developing products and services to test LTE Advanced Pro (LTE-A Pro) circuits and network equipment and even emulate entire networks in support of these test scenarios. This testing will provide valuable experience in the development of 5G test gear. Keysight s SystemVue 5G NR Library allows simulation of massive MIMO system designs early to fully characterize antenna end-to-end performance with embedded RF channel models before investing in expensive prototypes. With the move to mmwave spectrum, some tests will need over-the-air (OTA) validation. Keysight s Propsim channel emulation OTA testing solution enables verification of a device s performance in a real-world setting by accurately simulating the radio channel characteristics without needing to connect to a base station. Massive Machine Type Communications With the advent of mmtc, there will be a greater emphasis on connection density (up to 1000x that of current deployments), low power utilization (potentially 1% of current hardware), and deep coverage. High bandwidth and low latency are less of a consideration, and IPv6 support is a given. To a large extent, this use case is an extension of today s 2G/3G IoT networks, and much of the development will take place in designing optimized, low-cost, low-energy, and extended-life batteries and sensors. Ixia s XAir2 with high session counts, versus bandwidth per session, offers operators the first step in supporting different types of IoT tests. Keysight s battery analysis tools enables device optimization for power management with battery drain analysis test under different network conditions. Keysight s 5G protocol R&D toolset provides network emulation to verify designs with the many different network scenarios to ensure performance, coexistence, and collision free operation. Page 7

Ultra-Reliable and Low Latency Communications This challenging use case enables applications such as VR and autonomous control. It will depend on not only the New Radio described above, but also a new stateless software-defined architecture that includes distributed content and applications that reflect a SD-Test approach. Peak bandwidth requirements may not be that of enhanced mobile broadband, but this is balanced by high-speed mobility at up to 500 km/h (think of highspeed rail), low latency to 1 ms, and high reliability with potential packet loss of only 1 in 100 million. This is a segment that also mandates the strongest security guarantees. Low latency is partially implemented through a new feature in 5G NR called flexible numerology. With flexible numerology, slot duration time will scale to support the many different services expected in 5G. As noted earlier, URLLC will be defined after embb. Keysight is partnering with industry leaders to understand the requirements of this use model and to develop test hardware and software solutions that consider velocity, reliability, and latency. The test architecture must also correctly model the software-defined network architecture, and by extension, SD-Test. The final air interface, while it is still to be determined, will include orthogonal frequencydivision multiplexing (OFDM)-based waveforms that can support very large carrier bandwidths to enable the embb test cases. In addition, low latency will also require several baseband components to be optimized and a very close-to-the-air interface. Finally, while IoT will start with LTE -A Pro, it will eventually require 5G to support the massive scale and diverse nature of the devices. Keysight s 5G NR waveform generation and analysis tools enable early characterization of devices with new complex waveform structures that include flexible numerologies, bandwidth parts, uplink/downlink control, carrier aggregation and other 5G NR features that will enable future URLLC use cases. Designers are able to gain early insights into devices performance before the standards are complete. Page 8

Conclusion It is obvious that 5G is not going to have a monolithic, one-size-fits-all deployment, and any test architecture must reflect this. Based on the use cases, there are very different requirements with regards to bandwidth, density, latency, radio characteristics, energy use, and mobility. Given the phased deployment and target markets, 5G may, in fact, end up as a set of service offerings, sharing some technologies but differing widely in others. Keysight products and solutions are well positioned with LTE-A Pro and for the move to 5G: from LTE-A IoT and running on unlicensed spectrum, to increasing throughput in 5G from a combination of carrier aggregation, MIMO, and higher-order QAMs. Our products and solutions address the design and test needs across the ecosystem and through every layer of the protocol stack so that device designers, NEMs and operators can innovate, transform and win in 5G. Learn more at: www.keysight.com For more information on Keysight Technologies products, applications or services, please contact your local Keysight office. The complete list is available at: www.keysight.com/find/contactus This information is subject to change without notice. Keysight Technologies, 2018, Published in USA, April 23, 2018, 5992-2921EN Page 9