The wireless edge transformation has begun and will realize 5G's full potential - @qualcomm_tech August 2018
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
August 2018 @qualcomm_tech The wireless edge transformation has begun and will realize 5G’s full potential
Leading mobile innovation for over 30 years Digitized mobile Redefined Transforming communications computing industries Analog Desktop to Connecting virtually everything to digital smartphones at the wireless edge Transforming how the world connects, computes and communicates 2
Making 5G a reality in 2019 Industry-leading 5G End-to end Commercial Commercial R&D interoperable system chipsets: networks global standards prototypes Qualcomm® and products and test beds Snapdragon™ X50 5G modem Qualcomm Snapdragon is a product of Qualcomm Technologies, Inc. and/or its subsidiaries. 3
Our technology inventions drove Rel-15 specifications Scalable OFDM- Flexible slot-based Advanced Massive Mobile based air interface framework channel coding MIMO mmWave Scalable OFDM Self-contained Multi-Edge LDPC and Reciprocity-based Beamforming numerology slot structure CRC-Aided Polar MU-MIMO and beam-tracking Address diverse spectrum, Low latency, URLLC, Support large data blocks, Large # of antennas to For extreme capacity deployments/services forward compatibility reliable control channel increase coverage/capacity and throughput Early R&D investments | Best-in-class prototypes | Fundamental contributions to 3GPP 4
World’s first 5G NR milestones led by Qualcomm World’s first interoperable World’s first interoperable Successful multi-band Interoperable 5G NR sub-6 5G NR interoperability Rel-15 5G NR trials 5G NR sub-6 GHz data 5G NR mmWave data 5G NR interoperability GHz & mmWave testing preparing for the based on Snapdragon connection connection testing connections with 5 vendors Chinese mass market X50 modem chipset November 2017 December 2017 February 2018 MWC 2018 June 2018 2H-2018 Driving the 5G ecosystem towards 2019 launches in collaboration with 18+ global mobile network operators and 20+ device manufacturers 5
Operators select Snapdragon X50 5G modem for 5G NR trials in 2018 Vodafone Group OEMs select X50 5G NR modem for global device launches in 2019 6
5G Modem family World’s first 5G NR multimode modems 2G/3G/4G/5G in a single chip Sub-6 + mmWave Premium-tier smartphones in 2019 7
Commercializing mmWave in a smartphone form factor 76 mm 157.25 mm mmWave (60 GHz) viability Qualcomm® 5G NR 9.7 mm in handset form factor mmWave prototype 11ad in Asus 5G NR mmWave Zenfone 4 Pro Qualcomm® Reference Design Qualcomm Reference Design is a product of Qualcomm Technologies, Inc. and/or its subsidiaries. 8
Qualcomm® Smartphone form factor Fully-integrated Newly supported mmWave RF mmWave bands Advanced mobility features QTM052 mmWave antenna modules Suitable for compact Including transceiver, Supporting for up to 800 Supporting beamforming, smartphone industrial PMIC, RF front-end MHz of bandwidth in beam steering, and beam designs with four components, and a n257, n260, and n261 tracking for bi-directional Making mobile mmWave mmWave modules phased antenna array 5G NR mmWave bands mmWave communications viable and ready for commercialization Qualcomm QTM052 is a product of Qualcomm Technologies, Inc. and/or its subsidiaries. 9
Realize low latency Privacy New experiences Immediacy Edge cloud Customization To scale, Efficiency On-device intelligence must be distributed to the wireless edge, on the device Extract local value Reliability Trillions of connected things Massive amount of data Security and privacy Content closer to user Personalization 11
Cameras Sensors Mobile computing Automotive Mobile is becoming the Smartphones pervasive AI platform bringing broader Healthcare Smart homes Societal benefits Extended reality and the promise of massive on-device AI Wearables Robotics Rapid Superior Power/thermal Integrated/ replacement scale efficiency optimized cycles technologies Gadgets Industrial IoT Driver Virtual Enhanced Superior Healthcare safety assistant security photography and more…12
On-device intelligence is paramount Privacy Process data closest to the source, complement the cloud Perceive Hear, see, monitor, and observe Reliability Reason Immediacy Learn, infer context, and anticipate Training Execution/inference Efficiency Act Intuitively, interact naturally, protect privacy Execution/inference Training (emerging) Personalization 13
The challenge Constrained of AI workloads mobile environment Compute Thermally efficient intensive for sleek designs Large, complicated models Complex Requires long battery concurrencies life for all-day use Real-time Always-on Storage/memory bandwidth limitations Power and thermal efficiency Critical to the promise of AI on a wide range of connected devices 14
Security Perceive Hear, see, monitor, Sensing Our technologies and observe are transforming the wireless edge today Processing Reason Learn, infer context, and anticipate Inventing technology at scale to realize Connectivity the promise of massive on-device AI Providing the connectivity fabric with Act LTE, Wi-Fi, Bluetooth today Edge services Act intuitively, interact naturally, and protect privacy On-device AI 15
A unifying connectivity fabric for society Scalable to extreme simplicity Like electricity, you will just expect it everywhere 16
Leading society to 5G Rel-15 Rel-16+ And the expansion of the 5G ecosystem 17
LTE Advanced Pro accelerates the ecosystem expansion Drone communications 5G NR mmWave Public safety nx10 Gigabit 5G 5G NR Sub-6 GHz, Gigabit LTE and LTE coverage Digital broadcast nx1 Gigabit 5G VoLTE LTE IoT Private LTE networks Ubiquitous LTE Automotive Gigabit LTE, VoLTE (C-V2X) 18
A unifying connectivity fabric for the wireless edge Connecting virtually everything, an innovation platform for future services Extreme mobile broadband Mission-critical services Massive Internet of Things 19
Low latency multi- player gaming Distributed processing for boundless XR The wireless edge Instant cloud access transformation realizes On-device intelligence the full potential of 5G to act with immediacy Local interactive New experiences with new levels of immersion, content immediacy, personalization and privacy Creating new industries and transforming existing industries in the new era of distributed autonomy Edge sensors with privacy Essential on-device capabilities augmented with processing/compute, content, control,… at edge cloud Ultra-low latency for reconfigurable factory Virtual assistant with personalization 20
Enhancing services, creating new services, new Realize 5G’s deployment models with our end-to-end expertise low latency Augmented by Driving the best possible Distributed/split functions over 5G Cloud edge cloud on-device capabilities Big data, AI training, less delay Compute/processing, content, Sensing, processing, sensitive content, storage,… control, storage,.. closer to user1 security, intelligence Extract value – vs. ‘dumb’ data pipe Deeper On-premise in network factory/venue Economic and performance tradeoffs Customization, local value Latency could be Latency as low over 100s ms today as 1 ms Cooperation between operators, infra vendors, OEMs, content providers, cloud providers,… New deployments, like private networks 1. Such as distributed/virtualized core, distributed packet gateway functionality for low latency, mobile edge compute, related to MEC Multi Access Edge Computing as defined by ETSI 21
Distribute functions based on economic and performance tradeoffs for use case Driving the best possible on-device capabilities Cloud Edge cloud Big data, AI training, less delay Compute/processing, content, sensitive content, storage,… control, storage,.. closer to user1 On-premise control for On-car intelligence for ultra-low latency in the immediacy, edge cloud reconfigurable factory e.g. for local HD maps Deeper in network Economic and performance tradeoffs On-premise factory/venue Distribute processing ? New services through across device and edge distributed/split functions cloud for boundless XR over 5G Latency could be Distribute functions depending on use Latency as low over 100s ms today case e.g. to reach required latency as 1 ms Content/control closer to On-device virtual assistant for 1. Such as distributed/virtualized core, distributed packet gateway functionality for low latency, user for instant access, real- privacy and personalization— 22 mobile edge compute, related to MEC Multi Access Edge Computing as defined by ETSI time online gaming, etc. assistance from cloud
Instant access to media Connected cloud Low latency online The basics: responsive and entertainment computing multiplayer gaming web browsing Interactive venue experiences Rich real-time user-generated Mobile immersive Distributed processing for like live 360° streaming content, like video sharing experiences boundless photorealistic XR • Fiber-like data speeds • Content/control closer to user Enhanced and • Low latency • Realization of low latency • Uniform performance + • Customized local value = entirely new • Massive capacity • Augment on-device processing experiences 23 23
5G’s capacity and latency for shared venue experiences 24
Augment on-device processing for boundless photorealistic mobile XR 25
A glimpse into the future — sleek and stylish XR glasses How do we get there? Bone conduction transducers Multimode connectivity (4G, 5G, etc.) Directional speakers Many passive and active cameras with fisheye and telephoto lenses Optoelectronic night vision and thermal imaging sensors Ambient light sensors Tracking and recording cameras Eye tracking cameras Inertial, haptic, and health sensors New optics and projection Multiple high sensitivity technologies within a durable, audio microphones semitransparent display 26
Boundless mobile XR — experience the best XR anywhere On-device processing, On-device processing access to rich content augmented by edge cloud Premium Photorealistic graphics XR anywhere and visuals Efficient on-device processing to Enhanced experience where possible with deliver immersive XR new split-rendering architecture Utilize connectivity for less On-device processing augmented by compute time-sensitive content and downloads located at cloud edge over 5G connectivity We are doing this today We are working with the ecosystem to enable this 27
A new era in distributed processing Thermally efficient for Compute intensive sleek and ultra-light Complex concurrencies Complex Challenging XR workloads XR form factor Long battery life Real-time for all-day use Always-on Storage/memory Latency sensitive bandwidth Essential on-device processing Split rendering Augment by edge cloud processing Optimized under strict power, Significant higher power thermal, size constraints envelope—beyond PC class Premium experiences today that Augment on-device rendering with continuously improve edge cloud rendering Low latency High capacity Reliable link 28
Augment on-device processing for boundless photorealistic mobile XR XR headset Cloud Edge Cloud with on-device processing Less latency sensitive content with rendering (e.g. recorded streaming video) 6-DoF head pose On-device Encoded data adjustment to latest pose1 Deeper On-premise, in network e.g. venue Today’s latency is Partial rendering offload possible with 5G’s Power-efficient, latency sensitive unpredictable low latency, capacity and quality of service on-device rendering and tracking 1. Asynchronous time warp reduces Motion to Photon (MTP) latency by using on-device processing based on the latest available pose. MTP below 20 ms generally avoids discomfort — has to be processed on the device 29
AI offers enhanced experiences and new capabilities for smartphones True personal assistance Superior photography Extended battery life Natural user interfaces Enhanced connectivity Enhanced security A new development paradigm where things repeatedly improve 30
v Voice is the transformative user interface (UI) we’ve been waiting for Designed to be Always-on Conversational Personal Private Critical to create a true virtual assistant A context aware and personalized “digital me” sitting on the device 31
A true personal assistant is responsive and proactive Responsive Proactive Based on contextual analysis and prompting Based on contextual analysis without prompting (e.g. finding memories) (e.g. automatically sharing memories) “Remember the time I was strolling with “I noticed that you are tired and stressed, I’m turning my son after the party at La Jolla beach?” on the Rocky III soundtrack and navigating you to the gym for a workout and sauna.” “Yes I do, here is a picture you took of the sunset. “This music gets my blood going and a Should I share it with your family group on WeChat?” workout and sauna will help me relieve stress.” 32
Contextual intelligence required for personalization On-device data Off-device data Sensor data Calendar Cloud data/rich content Environment Iris scan Ambient light Compass Messaging IoT data Camera Microphone Temperature Humidity Sensor fusion Apps Gyroscope Pulse C-V2X The local fusion of many types of sensors and personal information 33
On-device AI creates a true personal virtual assistant Cloud-centric (today) less privacy and personalization Automatic speech Voice recognition AI agent activation Service SMS Maps Natural language manager Weather News understanding SensorEcho cancel, Video Music Text-to-speech fusion speech denoising Access to rich content On-device (future) Continuously learns personal Sensor fusion knowledge and acts intuitively Automatic speech Voice recognition activation with immediacy Service SMS Maps Natural language understanding manager Weather News AI Echo cancel, Video Music Text-to-speech agent speech denoising 34 Both ends are needed — but AI functions gradually moving on to the device
Power efficient AI Personalized AI Qualcomm Technologies is making on-device Model design Model quantization Contextual Self- AI ubiquitous learning Efficient hardware Model Model Privacy/ Low-power activation compression Algorithmic advancements security always-on Software tools 35
All data should not go to the central cloud—especially private Cloud based processing Security camera Notification Storage Raw video stream Camera processing, deep learning Wireless edge processing Security camera Relevant video clips Aggregated edge data analytics Notification On-device camera processing, deep learning Lower latency, privacy, on-device security, but also less bandwidth 36
AI in here. And everywhere. Advanced camera processing, powerful machine learning, and computer vision at the Refrigerator Inventory edge will allow new applications that push the No eggs available Chicken boundaries of our connected world. Bell Peppers Soy Sauce Milk level is low Purple Basil Raspberries Milk (level is low) MENU SUGGESTIONS Intelligent cameras Local at the edge monitor and track processing means what’s inside. low latency and max efficiency. 2 pizzas available Dish soap level is low Neural networks Ubiquitous connectivity watch and learn helps you avoid running your preferences. out of critical food supplies 37
7YTR843 3NDU913 783HNW K8925-A Alert! Minor collision in area. First responders are on the scene. 2LBA642 7TOZ033 7JPS440 5ELP440 5KTZ714 Intelligent License Plates visible from my position: 1XAS876, 7JPS440, 783HNW, 5KTZ714, 3MNS337, 5LBR363, 1ILM281, Smart Cities 27IWS291, GRT-BRTN, 3UJS301, 6SAW391, 1NAL963, 5ELP440, 1XAS876 7TOZ033, 2LBA642, K8925-A, 3NDU913, 3ANL538, 7YTR843… • Scene Classification • License Plate accidentRecognition Traffic slow ahead due to an in the left lane. • Target Sound Detection 5LBR363 3MNS337 1XAS876 1ILM281 27IWS291 3UJS301 1NAL963 GRT-BRTN 6SAW391 38
Massive scale Power efficient To efficiently support dense To realize10+ year device battery life and 100x network energy 2 connections of 1+ million efficiency 3 devices/km 2 Scaling for the massive Internet of Things In-band 5G NR (R16+) LTE IoT Continued eMTC evolution Long range Extreme simplicity To reach challenging locations To allow scaling to the lowest-end Today Continued NB-IoT evolution by achieving device link budget use cases with e.g., single Rx of 164 dB1 antenna Addressing growing needs of low- power, wide-area IoT use cases 1. Maximum Coupling Loss, assuming data rate of 160bps 2. Assuming 200B UL + 20B DL per day at 164 MCL with 5Wh battery 3. Compared to IMT-Advanced 39
5G IoT requires new solutions at the wireless edge Opportunities Challenges Our solutions Trust environment Device-cloud rooted in HW integration at scale Monetize and manage massive amount Trust and security of low-complexity things Large number of ecosystem participants Manage new, more intelligent devices and new use cases New business and deployment model, Total cost of ownership like private networks for the Industrial IoT HW/SW, life-cycle support, upgradeability Zero-touch device life-cycle management Trusted chipset services for enterprise and IoT providers 40
Trusted services to securely connect and manage devices Qualcomm® wireless edge services cloud platform Unique device key embedded in chipset at factory • Secure boot, debug • Hardware root-of-trust • Secure execution environment • Runtime integrity checks 3rd party device Secure and trusted life cycle management management platform • Device attestation and connection integrity • Device provisioning with plug-n-play on-boarding • On-demand chipset upgrades and feature activation 41 Qualcomm wireless edge services is a product of Qualcomm Technologies, Inc. and/or its subsidiaries.
Enabler to the factory Safer, autonomous Reliable access Precision of the future transportation to remote healthcare agriculture Efficient use of Private networks for logistics, Sustainable smart cities Digitized logistics energy and utilities enterprises, industrial,… and infrastructure and retail Powering the digital economy 5G will expand the mobile ecosystem to new industries > $12 Trillion In goods and services by 2035 * 42 * The 5G Economy, an independent study from IHS Markit, Penn Schoen Berland and Berkeley Research Group, commissioned by Qualcomm
Driving a rich 5G roadmap in Release 16 and beyond 5G massive 5G NR in unlicensed/shared IoT spectrum 5G NR 5G NR integrated C-V2X access and backhaul 5G industrial IoT 3GPP Rel-15 5G broadcast with URLLC design provides the foundation for R16+ Sub-6 GHz | mmWave 43
Private 5G networks for Industrial IoT use cases Optimizing LTE today New opportunities with 5G NR Ultra-reliable Time-sensitive low-latency networking Scalable from Gigabit LTE to LTE IoT mmWave and Wireline ethernet sub-6GHz eMBB replacement Optimized Dedicated Secure Tailored for industrial applications, Local ‘edge’ network, easy to Industrial grade security e.g., QoS, latency deploy, independently managed with LTE and 5G NR 44
Container ports Oil refineries Manufacturing Construction Mines Warehouses Wind farms Oil rigs > $5 Trillion 1 Global economic output in 2035 enabled Manufacturing $3,364B Transport $659B Construction $742B Utilities $273B Mining $249B by 5G in the following five categories 1. “The 5G economy: How 5G technology will contribute to the global economy” by IHS Economics / IHS Technology 45
Enabling reliable wireless IoT connectivity at transport hubs 46
The reconfigurable factory of the future will thrive on the wireless edge 47
Security camera Sensors Extreme Massive IoT mobile broadband Process Monitoring Latency: 100 ms Availability: 99.99% Rate: kbps Head mounted display Augmented Reality Latency:
Cloud services Local network On-device Cloud analytics and virtualized Local gateway and/or local Various degree of Sensor core network functions core network functions on-devices capabilities fusion Sensors Security Gateway 15.4 W-HART Analytics Wi - Fi BLE Computer vision Wired Machines Sensitive traffic/data stays on site Enables ultra-low On-device 1 ms latency intelligence Access Big data analytics Edge analytics Low-power processing 49
You don’t want your device to confirm with the cloud to take action On-device intelligence is key for today’s and future autonomous cars 50
On-device intelligence is key for autonomous cars Autonomous cars Perceive Reason Act Vision Radar Tele-operation Machine learning Sensor processing Automotive SoC Safety MCU Data analytics Precise algorithms Machine learning and SW HD maps positioning Sensor fusion Fail-operational compute Mission planning Drive-by-wire Lidar Behavioral analysis vehicle controls Edge cloud for local Central cloud for C-V2X data, e.g. HD maps big data/training 51
Communicating directly with people and things V2I V2V C-V2X Car intends Establishes the foundation Green for 20s to enter lane V2N for safety use cases and Rerouting a continued 5G NR C-V2X evolution for future Construction 1.3 mi autonomous vehicles Release 14 C-V2X completed in 2017 V2V Non-line of sight Car approaching Broad industry support, from blind corner 5GAA Global trials started in 2017 V2P Cyclist approaching Our 1st announced C-V2X product in September, 2017 52
C-V2X R14 only car 5G NR C-V2X R14 C-V2X brings non-line of sight safety capabilities Brings complementary capabilities for autonomous driving Path planning Intention and trajectory sharing for faster, yet safe maneuvers C-V2X R14 / R16 car Share perception Sharing of high throughput sensor data and real world model Real-time updates Real-time sharing of local data with infrastructure and vehicles (e.g. 3D HD maps) Coordinated driving Exchanging intention and sensor data 53
Our technologies are already transforming the wireless edge to realize the full potential of 5G Learn more at www.qualcomm.com / 5G 54
Thank you! Follow us on: For more information, visit us at: www.qualcomm.com & www.qualcomm.com/blog Nothing in these materials is an offer to sell any of the References in this presentation to “Qualcomm” may mean Qualcomm components or devices referenced herein. Incorporated, Qualcomm Technologies, Inc., and/or other subsidiaries or business units within the Qualcomm corporate structure, as ©2018 Qualcomm Technologies, Inc. and/or its affiliated applicable. Qualcomm Incorporated includes Qualcomm’s licensing companies. All Rights Reserved. business, QTL, and the vast majority of its patent portfolio. Qualcomm Qualcomm and Snapdragon are trademarks of Qualcomm Technologies, Inc., a wholly-owned subsidiary of Qualcomm Incorporated, registered in the United States and other Incorporated, operates, along with its subsidiaries, substantially all of countries. Other products and brand names may be Qualcomm’s engineering, research and development functions, and trademarks or registered trademarks of their respective substantially all of its product and services businesses, including its owners. semiconductor business, QCT.
You can also read