Designing ultra-low latency 5g applications
- 1. DESIGNING ULTRA-LOW LATENCY 5G APPLICATIONS GABRIEL PAREDES LOZA May 2024 AWS Expert
- 2. 2 THE SPEAKER Gabriel Paredes Loza Gabriel has 10+ years of AWS Cloud experience working in a variety of roles and companies. With big passion for Data & AI, IoT, DevOps, and Cybersecurity, he is on a mission to fill the gap between digital innovations and business outcomes.
- 3. 3 Ultra-Low Latency Applications Implementation Challenges AWS & 5G Conclusions Q&A AGENDA 07.05.2024 / Designing Ultra-Low Latency 5G Applications
- 5. 5G allows three service grades that may be tuned for key requirements, allowing customers to validate different business cases. 5 ULTRA-LOW LATENCY APPLICATIONS 5G’s Service Grades Enhanced Mobile Broadband (eMBB) Ultra-Reliable-Low Latency Communication (URLLC) Massive Machine Type Communication (mMTC) Gigabits in a second Smart Home / Building Voice Smart City 3D Video, 8K, UHD Screens Cloud Games and Work Augmented Reality Industry Automation Self Driving Cars Mission Critical Apps
- 6. 6 Ultra-Low Latency Applications require more than “just” ultra-low latency… 5G Feature Name 5G Feature Description 5G Feature Specification 4G Feature Specification Broadband Speed Maximum Data Transmission Rate per User across Coverage Area 20 Gbit/s to 100 Mbit/s 100 Mbit/s to 10 to 15 Mbit/s Latency Time to transmit a Data Packer from one Node to Another 1-5 milliseconds 5-50 milliseconds Stable Performance Under High-Speed Mobility Conditions Quality Handover Interruption Time, Uplink, Downlink, and Jitter Quality up to 500 km/h Quality up to 350 km/h Connection Density Ability to support Data Delivery within a certain time in a space- constrained location Over 100.000 Connections/km2 Around 2.000 Connections/km2 Note: all data points are approximations, given the existence of many different standard specifications. Source: The CIO’s Guide to 5G Connectivity and Strategy, Forrester Report ULTRA-LOW LATENCY APPLICATIONS
- 7. Required Investments for 5G Networks Rollout are backed by several Business Drivers the technology’s Value Proposition Unlocks. 7 ULTRA-LOW LATENCY APPLICATIONS Definition & Drivers High Bandwidth Low / Ultra-Low Latency Quality of Service Long-Range Coverage Long Battery Life Real-Time Decision Making Improved Efficiency & Productivity Enhanced Customer Experience Costs Savings New Business Streams / Models Business Drivers 5G’s Value Proposition Unlocks Requires
- 8. 8 5G Connected Products 5G infrastructure can connect tens of millions of wireless devices and enables M2M communication links in IoT 5G Networks as a core enabler for IoT, AI/ML, and Edge Computing use-cases AI / ML 5G enables efficient deployment of key automation use-cases leveraging AI / ML 5G Unlocks Value by leveraging complementary technologies along the value chain Edge 5G enables data processing close to the connected device, using ultra-low latency and ultra-high bandwidth Source: “5G Part Three: IoT, Edge Computing, and Network” Webinar, Forrester Report. ULTRA-LOW LATENCY APPLICATIONS
- 10. Components Failure/ Disruption Increased Devices & Data Volumes Larger Bandwidth Network Shared Network (Resources) Different Traffic (Categories) Vast # Consumers & Producers Vastly Expanded Attack Surface Data Privacy & Governance Vastly Expanded Attack Surface 10 IMPLEMENTATION CHALLENGES Technical Considerations Cybersecurity Data Governance Quality of Service Scalability & Resiliency APIs / Entry Points
- 11. 11 11 REFERENCE ARCHITECTURE Building Blocks: Devices & Radio Interface O-RAN Empowered by principles of intelligence and openness, the O-RAN Architecture is the foundation for building a virtualized RAN on open hardware and cloud, with embedded AI-powered radio control.
- 12. 12 12 REFERENCE ARCHITECTURE Building Blocks: Edge Infrastructure Multiple-Access Edge Computing (MEC) Bring logic and cloud computing power to the edge of a network, closer to end users, reducing latency, and network congestion.
- 13. 13 13 REFERENCE ARCHITECTURE Building Blocks: Core Network & Internet, Cloud Infrastructure Distributed Cloud Public cloud services available in remote locations, a single control plane for unified developer / strategy experience, and secure & reliable network tunnels, are the key components of a distributed cloud implementation.
- 14. 14 14 REFERENCE ARCHITECTURE Building Blocks: Edge-to-Cloud Governance Platform Governance E2E Data Security & Governance API Layer Definition
- 15. ❗ Unreliable Network Performance ❗ Latency Values in the 100ms range ❗ Network Backbone Data Pressure (e. g. 10Gbit/s with 5G) 15 REFERENCE ARCHITECTURE Deployment Type: Cloud-Based AI
- 16. ü Ultra-Low Latency UE <-> Edge (5ms RTT)* ü Data Pressure terminated @ Edge) ❗ Unreliable Network Performance Edge <-> Cloud 16 REFERENCE ARCHITECTURE Deployment Type : Edge AI *: Latency values are experimental and have been measured by using AWS Wavelength as MEC.
- 17. ü Service Link Edge <-> Cloud ü Low Latency UE <-> Cloud (<50ms RTT)* ü Single Control Plane for Cloud and Edge 17 REFERENCE ARCHITECTURE Deployment Type : Distributed AI *: Latency values are experimental and have been measured by using AWS Wavelength as MEC and AWS PrivateLink in the Parent Region’s VPC.
- 18. AWS & 5G
- 19. 19 AWS & 5G 5G Offering on AWS* *: Some services might or might not be available outside the US. Please consult the AWS website before planning their usage. AWS Wavelength Run compute and storage workloads on AWS infrastructure located on the edge of the 5G network AWS Outposts Extend AWS infrastructure, services, APIs, and tools to on-premises data centers and edge locations AWS Private 5G Deploy, operate, and scale own Private Mobile Network, with all needed HW and SW provided by AWS AWS Local Zones Place compute, storage, database, and other selected AWS services close to large populations or industry centers
- 20. 20 AWS & 5G AWS Wavelength
- 21. 21 AWS & 5G AWS Wavelength
- 22. 22 AWS & 5G AWS Wavelength
- 23. 23 AWS & 5G AWS Wavelength
- 24. 24 AWS & 5G AWS Wavelength
- 25. 25 AWS & 5G AWS Wavelength
- 26. 26 AWS & 5G Ultra-Low Latency AWS Reference Architecture ü Cloud-Native ü Highly Available ü ML Inference ≦ 10ms
- 27. 27 AWS & 5G Ultra-Low Latency Use-Cases & Focus Industries 100s of Gbps 100s of Gbps 100s of Gbps 100s of Gbps 100s of Gbps 10s of Gbps 10s of Gbps 10s of Gbps 10s of Gbps 10s of Gbps
- 28. CONCLUSIONS
- 29. 29 CONCLUSIONS “Ultra-Low Latency Applications require more than “just” Low Latency” Key Takeaways ！ 5G’s Future lays on Justified ROIs involving significant Implementation and Modernization Costs ！ Scaling-up ULL Solutions generates challenges such as Increased Data Volumes and Attack Surfaces ü Distributed Cloud with AWS provides a Single Control Plane for the setup and maintenance of an E2E solution ü Edge AI, 5G, and IoT are the perfect Triade for Unlocking new ULL Use-Cases and Business Opportunities
- 30. THANK YOU! Gabriel Paredes Loza AWS Expert