Public Cloud Edge Interface (PCEI) BP Family: Public Cloud Edge Interface Target Industry: IoT, Developers, Operators, Clouds, DCs, Interconnection Purpose/Features: The purpose of Public Cloud Edge Interface (PCEI) Blueprint is to specify a set of open APIs and orchestration functionalities for enabling Multi- Domain Inter-working across functional domains that provide Edge capabilities/applications and require close cooperation between the Mobile Edge, the Public Cloud Core and Edge, the 3rd-Party Edge functions as well as the underlying infrastructure such as Data Centers, Compute hardware and Networks Use cases & Applications ● Edge Multi-Cloud Orchestrator (EMCO) - PCEI Enabler ● Deployment of Azure IoT Edge Cloud Native PCE App ○ Using Azure IoT Edge Helm Charts provided by Microsoft ● Deployment of AWS Green Grass Core PCE App ○ Using AWS GGC Helm Charts provided by Akraino PCEI BP ● Deployment of PCEI Location API App ○ Using PCEI Location API Helm Charts provided by Akraino PCEI BP ● PCEI Location API Implementation based on ETSI MEC Location API Spec ● Simulated IoT Client Code for end-to-end validation of Azure IoT Edge ● Azure IoT Edge Custom Software Module Code for end-to-end validation of Azure IoT Edge NEW!
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Purpose/Features:The purpose of Public Cloud Edge Interface (PCEI) Blueprint is to specify a set of open APIs and orchestration functionalities for enabling Multi-Domain Inter-working across functional domains that provide Edge capabilities/applications and require close cooperation between the Mobile Edge, the Public Cloud Core and Edge, the 3rd-Party Edge functions as well as the underlying infrastructure such as Data Centers, Compute hardware and Networks
● LF Edge’s Project EVE-OS to provide remote management, Zero
Trust security (physical and software)● LF Edge’s Fledge as an IIoT framework for sensors, historians, DCS,
PLC’s, and SCADA systems and connectivity to public or private clouds
● Remote monitoring and updating of applications, without bricking the device
● AI Models, real time data capture, and cleansing at the device edge● Sample application that can be customized to meet many different
Use Cases
Use cases & Applications● Predictive Maintenance
● Hazards monitoring (People detection in hazardous area)
NEW!
The AI Edge: Federated ML Application at Edge
BP Family: AI Edge
Target Industry: Driverless cars, Warehouse
Purpose
To provide a Federated Learning Platform that trains Machine Learning algorithm across edge devices without them sharing the data that make up the models.
Features● FATE first unsupervised learning algorithm: Hetero KMeans● Add Data Split module: splitting data into train, validate, and test sets
Target Industry: Smart road, Cold chain logistics, Smart building, etc.
Purpose:
● First Release will focus on the ML inference offloading Use Case
Features:● KubeEdge managed Application
deployment and life cycle management● ML offloading to Edge server● Cloud(training), Edge (Inference), Device
collaboration
NEW!
Kubernetes Native Infrastructure for Industrial Edge
BP Family: KNI
Purpose/Features:● Managing edge computing clusters from a central
management hub by using Advanced Cluster Manager● GititOps based application deployment with ArgoCD● Cloud Native CI/CD Pipelines with Tekton● Event streaming from edge to core with Kafka AMQ
Streams and Mirror Maker● Machine learning as a data scientist with Jupyter Notebook.
Use cases & Applications● Machine inference-based anomaly detection
NEW!
Target Industry: Manufacturing
The AI Edge: Intelligent Vehicle-Infrastructure Cooperation System(I-VICS)
BP Family: AI Edge
Target Industry: Autonomous Vehicles
Purpose/Features:
● Autonomous Valet Parking
Use cases & Applications● Starting and testing the behavior planner● Starting and testing the global planner● Initializing the NDT localizer● Running the EKF filter for localization● Trajectory Following
Creating a EPC/5G “in a box” to enable enterprises and operators to deploy LTE/5GUses OSS such as free5GC/Magma
NEW!
Enterprise Application on Light weight 5G Telco Edge (EALTEdge)
BP Family: 5G MEC/Slice
Target Industry: Telco operators
Purpose/Features:Provides a complete ecosystem for enterprise applications on light weight 5G Telco Edge. Can be leveraged by Telco operators to provide edge computing capability to it’s enterprise users. Overall objective of this blueprint is to provide the following main features.
R4 Improvements● Leverage EdgeGallery to add application/MEC Edge Orchestrator, Dev Platform,
Dev and Tenant Portals● Built a sample ROBO
Use cases:ROBO(Remote office Branch office): Due to limited resource and disaster prone of ROBO sites, edge native storage, Backup and restore on lightweight telco edge is supported. Smart retail with automatic shelf management on ROBO sites is developed and integrated. Machine Vision on Campus Networks: Centralized processing using wireless cameras, real-time response for detection/feedback; provide shared GPU
5G/MEC Slice System to Support Cloud Gaming, HD Video & Live Broadcast
BP Family: 5G MEC/Slice
Target Industry: Gaming, Video, Broadcast
Purpose/Features:The 5G MEC BP consists of two network elements. One is the edge connector which is deployed in the cloud to enable traffic offloading, subscribe edge slice and implement application lifecycle management etc. The other is the edge gateway which is deployed close to the 4G/5G network to perform traffic steering, Local DNS service and traffic management etc.
Use cases & Applications● Cloud Gaming● HD Video● Live Broadcasting● Small deployment targeting MEC in access sites or
enterprise● Medium deployment targeting MEC in central offices
Purpose/Features:Enables new functionalities & business models on network edge. Benefits include better latencies for end users; less load on network, since more data can be processed locally; and better security and privacy, since sensitive data need not be transferred to a centralized location.Use cases:● Fixed installation as part of 5G NR base stations; enables new services that
leverage especially low latency, such as AR/VR● As an extension of the previous, the “Smart City” deployments have additional
functions such as weather stations, cameras, displays, or drone charging stations. The control software for these functions would run on the uMEC
● In an Industry 4.0 use case set, the uMEC is deployed as part of a 5G network and would provide a platform for running services for the factory floor
● In a train, the uMEC can collect and store surveillance camera data for later uploading
AI Edge: School/Education Video Security MonitoringBP Family: AI Edge
Target Industry: Education, Home
Purpose/Features:Focuses on establishing an open source MEC platform combined with AI capacities at the Edge; can be used for safety, security, and surveillance sectors as well as Intelligent Vehicle-Infrastructure Cooperation Systems.
Use cases:● Hierarchical cluster management● Duplex channel between cloud center and
edge cluster● Kubernetes native support ● Accurate routing of messages between clusters● Support both x86 and arm64
IEC Type 3: Arm-Enabled Android Cloud Applications
BP Family: IEC
Target Industry: Gaming
Purpose/Features:Supports Android applications and services running on Arm-enabled cloud architectures with GPU/vGPU EC management. Arm-based- cloud games need basic “cloud” features, such as flexibility and broad availability, which this blueprint provides.
R4-● Android application environment based on Robox ● GPU Support
Use cases:
● Android Cloud Games: compress the rendering of game scenes into video and audio streams on the edge Android platform.
Then edge cloud server transmits the compressed game pictures to the players' game terminals through a 5G network, and obtains the
players' input instructions to realize interaction. End to end latency better =< 20ms.
● AR/VR Android Applications
IEC Type 5: SmartNIC
BP Family: IEC
Target Industry: Telco and other carriers
Purpose/Features:IEC Type 5 is focused on SmartNIC, which can accelerate network performance and provide more management convenience.In general, the architecture consists of two layers: IaaS (IEC), SmartNIC layer. But in R3, we have two simple layers: Host Layer, SmartNIC Layer.
Use cases:● OVS-DPDK offload into SmartNic: accelerates
network performance and saves computing resources
● Part of the UPF and VPC functions, like load balancing, forwarding, dpi, etc offloaded into SmartNIC to enhance the performance of UPF that will be deployed in carrier's edge cloud datacenters
Purpose/Features:• Addresses IOT & Universal CPE use case• Targets IOT Appliances• Very thin OS and Orchestration• Full CI/CD deployment ready and verified• Platform is ready to support different IOT Gateway use
cases for Edge computing. Video Analytics is one of use case verified on this platform.
Updates in R2:• Integrated EdgeX framework for IIOT
• Supported/verified on Tailored OS, Ubuntu and CentOS• Single- click installation
• Portal for IOTgateway or uCPE with enabled features like application and platform management
• Enables community validation testing in CI for Hardware, OS and K8s layers.
Public Cloud Edge Interface (PCEI) BlueprintPCEI blueprint pursues development of multi-domain interworking capabilities to enable Mobile Operators, Public Clouds Core and Edge Compute providers as well as 3rd-Party Edge Compute providers utilize distributed data center infrastructure, interconnection and edge services for mobile edge cloud use cases such as Mobile Hybrid/Multi-Cloud, Multi-MEC access. ○ Joined PCEI blueprint as Project Technical
Lead○ Proposed PCEI Reference Architecture○ Participated in the development of first PCEI
feature based on OMA Zonal Presence API / ETSI MEC Location API
○ Lead development and implementation of PCEI for Akraino Release 4 demonstrating EMCO orchestrator and deployments of Public Cloud Edge apps from Azure and AWS
June 2020
POC & Deployment● SmartNic: In R3 provide the POD environment for
ByteDance, realized the offload of OVS-DPDK for SmartNiC, to increase the throughput of edge network VPC; In R4, China Mobile will provide a POD environment, with one BF card from Mallnox;
● Android: In R3, used ANBOX to deploy a containerized Android system on used an Arm-based server and conducted initial functional tests. Cooperated with ByteDance and Mozhiyun respectively to provide private Lab environment, implement CI/CD environment deployment in the private lab;
● PCEI: transplant ETSI MEC location APIs and will verify them in China Mobile private lab in China.
Community Contribution Focus● SmartNic: Focus on offloading network functions,
improving network throughput and enhancing management of network card resources.
● Android:Focus on the virtual deployment of Android cloud native applications on the Arm edge cloud.
● PCEI: Focus on provide the 5G core network functions to public cloud, improve the ETSI MEC APIs and build a unique API enabler between Telco and Cloud.
Lab resource: China Mobile provides MEC POD environment in Beijing for multiple BPs. 5G resources and accesses are under coordinating.
● KubeEdge Edge Service Blueprint
● This blueprint family showcases an end-to-end solution for edge services with KubeEdge centered edge stack. The first release will focus on the ML inference offloading use case.○ Initiated blueprint project ○ Proposed the Architecture○ Contributing to the
development of end-to-end lab validation environment
• Contributed to ELIOT: Edge Lightweight and IoT Blueprint Family project
• Open Source ONAP software company focusing on 5G/edge computing application automation
• New ONAP integration in the Akraino Private LTE/5G Blueprint
• Successfully completed 12 ONAP engagements
• Aarna Networks ONAP Distribution 4.0 (El Alto) available
• Recently joined PAWR, 5G Open Innovation Lab to drive 5G use cases with ONAP
• Number#1 Instructor led ONAP training provider
POC & DeploymentAI Edge supports video security monitoring, classroom concentration analysis, factory safety production, kitchen hygiene monitoring, and also scenarios in Intelligent Vehicle Infrastructure Cooperation System. In R3, cooperated with Arm, Intel, and Huawei, set up a private lab environment, implemented CI/CD environment. More AI application for Arm architecture will be released in the future.
Community Contribution FocusFocuses on establishing an MEC platform that combined with AI capacities at the Edge site. And it also could be used to enable the autonomous driving industry.
Enabled Arm architecture based hardware and software support for multiple blueprint families. These include several blueprints that share a similar set of use cases, software, and continuous integration and deployment.● Connected Vehicle Blueprint● Edge Lightweight and IoT (ELIOT)
Intel co-founded Akraino Edge Stack, continuously supported and contributed to the growth of the Edge ecosystem.
● Donated IA servers in Akraino Community Lab, plus supporting partners working on ICN and 5G MEC w/ Intel hosted PODs.
● Drove Integrated Cloud Native BP Family created SW Platforms for Enterprise, IoT and Telco markets, including MICN BP and Private 5G BP.
● Enabled Akraino R3 active community BPs with Intel architecture based hardware and software supported:○ 5G MEC Slice System to Support Cloud Gaming, HD Video and
Live Broadcasting BP○ Connected Vehicle BP○ Edge Lightweight and IoT (ELIOT) - ELIOT SD-WAN/WAN Edge/uCPE BP○ Kubernetes Native Infrastructure (KNI) – Provider Access Edge BP ○ The AI Edge - School/Education Video Security Monitoring BP
○ The AI Edge: Intelligent Vehicle-Infrastructure Cooperation System(I-VICS)
June 2020
• Juniper Network has been an active contributor in the Akraino community from the early days of its formation. They have been contributor for all three Akraino releases.
• Network Cloud with Tungsten Fabric BlueprintThis blueprint is part of release 3 which integrates Tungsten Fabric in Network Cloud. It integrates with Regional Controller to deploy edge sites that supports both Kubernetes as well OpenStack based workloads. Tungsten Fabric provides advanced networking SDN features to the edge sites.
• Juniper is also engaged with in the Akraino Private LTE/5G Blueprint
As part of Akraino R4, Huawei is associated with following blueprints family:
Enterprise Applications on Lightweight 5G Telco Edge : BP intends to provide an ecosystem for enterprise application on light weight 5G Telco Edge which can be leveraged by Telecom operators to its enterprise users. BP having following salient features:
● Lightweight MEC Solution with reference to ETSI MEC Architecture.
• Worked on validating the O-RAN Near-Real Time Radio Intelligent Controller (RIC) in a live network, using the Akraino REC project
• Promoted the emerging ETSI MEC ecosystem
• The RAN Intelligent Controller Project utilized the NokiaAirframe Open Edge Server Hardware that is based on Open Compute Project Design. Open Edge provides Ultra-small footprint for easy installation at the network edge; an extended temperature range, robust seismic tolerance enabling deployment worldwide; and provides the performance and low latency required by Cloud RAN and MEC.
• NN
June 2020
108
Akraino Commercial updates
1. POC & DeploymentThe AI Edge: Federated ML application at edge provide Federated Learning Platform for data stored locally, improves accuracy in the edge computing. FedVision is provided in R3. More federated applications and quick validations will be provided in the future release.
2. FedVisionA machine learning engineering platform to support the development of federated learning powered computer vision applications.
3. Community Contribution FocusFocuses on providing a federated learning platform which can be used in privacy protected and distributed edge applications such as vision, financial technology, Marketing Intelligence.
June 2020
POC & Deployment
●Connected Vehicle Blueprint can be flexibly deployed in physical machines, virtual machines, containers and other environments. TARS framework is an important open source component of Connected Vehicle Blueprint, which can efficiently complete the massive deployment and governance of micro-services.● IEC Type 4 AR/VR applications, in general, the architecture
consists of three layers: Iaas(IEC), PaaS(TARS), SaaS(AR/VR Application). TARS framework can efficiently complete the massive deployment and governance of micro-services, and make AR/VR applications deployed in physical machines, virtual machines, containers and other environments. ●5G MEC/Slice system to support cloud gaming, HD video and
live broadcasting: provides an edge connector and edge gateway to enable traffic offloading to edge applications, and supports application lifecycle management by using openNESS in R3. Means to subscribe edge slice, intelligent traffic management and enhanced local DNS will be provided in the future release.
Community Contribution Focus●Connected Vehicle Blueprint, focuses on Internet of Vehicles
(IoV) application MEC platform, which helps the rapid landing of IoV applications.● IEC Type 4 focuses on AR/VR applications running on edge.●PCEI: Focus on use the 5G MEC open API provided by
operator to support 5G MEC solution based on public cloud(i.e., ECM)
• Proof of Concept (PoC) completed for Akraino KNI R2 release on baremetal servers in 5G Lab.
• Proof of Concept (PoC) completed for Akraino KNI R3 release on virtual baremetal in 5G Lab.
• Implementing OpenAirInterface (OAI) use case on KNI R3.
• Showcasing Akraino and KNI blueprint to customers
Project Introduction: EdgeX
A highly flexible IOT open source software framework that facilitates integration and interoperability between heterogeneous devices and applications.
Top Use CasesEdgeX provides a common set of horizontal capabilities to support use cases across any IOT vertical. Examples:
› Manufacturing Remote monitoring of production equipment, get data from multiple sources and filter/transform it to react at edge before sending to the cloud for aggregation, analysis and to optimize production and maintenance.
› Retail - the Open Retail Initiative (ORI) promotes the EdgeX framework in retail to ingest data from cameras (OpenVino), POS systems,RFID, etc and use it at the edge for use cases like Loss Prevention and Inventory Management.
› Building Automation- Edge Control (control devices via a common API), use edge data to control building environment (HVAC, lighting, access). Connect to the cloud to optimize power consumption using ML.
Technical Summary› Agnostic to: HW, OS, OT protocols, sensors and Cloud & Enterprise endpoint› Distributable set of microservices for scalability and fault tolerance › Enables autonomous operations and intelligence moving to the edge to address
low latency decision making/actuation, bandwidth & storage, and remote ops
Project Release Status (as of Q2 2020)› 4 Million downloads! Growing at 1 million a month!› Geneva (1.2) commercial deployment release › Next release Hanoi (1.3)- Oct 2020- Improved Security,
● Developed by the SI Technotects after an end customer independently recognized the potential for EdgeX to provide them with more flexibility and reduce costly runtime licensing fees
● Leveraged Dell Edge Gateways, IOTech Edge Xpert, VMware Pulse IoT Center for management, Photon OS, RedisEdge for data persistence, and both Inductive Automation Ignition and native EdgeX Device Services for data ingestion
● Technotects was able to successfully reproduce the customer’s use case with EdgeX while also providing them with more options from the open ecosystem
● See the case study blog for more detail
Block diagram for the PoC
Water treatment
skid monitored as part of the
PoC
EdgeX in Smart Cities (Water Treatment)
EdgeX in Commerce – Intel RRK for theOpen Retail Initiative (ORI)
114
AutonomousVehicles
…
Enterprises Public buildings
Other Telco Real states
(Wire Centers, etc.)
● EdgeX serves as the foundation for the Intel-led Commerce Working Group within the EdgeX project
● Related Intel RFP Ready KiT (RRK) includes content from IOTech (Edge Xpert), Pixeom (container orchestration), HPI and Dell (edge HW) in addition to linking to Intel’s OpenVINO computer vision framework
● Video events from OpenVINO ingested into EdgeX for analysis in concert with telemetry from other sensors (e.g. building and energy systems, RFID)
EdgeX in the OMPAI AI Testbed within the IIC
115
AutonomousVehicles
…
Enterprises Public buildings
Other Telco Real states
(Wire Centers, etc.)
● Within the Industrial Internet Consortium (IIC), the Optimizing Manufacturing Processes by Artificial Intelligence (OPMAI) testbed leverages ML algorithms, technologies, and technical frameworks to apply optimally distribute AI from edge to cloud to solve specific production quality, cost & efficiency problems in an automotive manufacturing environment.
● Leveraging the EdgeX framework, AI models and edge applications are run for the local optimization of manufacturing processes. In the cloud platform, they are run to enable global and long-term optimization, e.g. across production lines and plants.
› Increase modularity to support more deployment options› Add support for Kubernetes via K3S and clustering › Add mesh networking capabilities› Continue to shrink the footprint of EVE in order to run on
smaller and resource-constrained embedded edge devices
https://www.lfedge.org/projects/eve/
An open edge computing engine that simplifies the development, orchestration and security of cloud-native applications on distributed edge hardware. Supporting containers, VMs and unikernels, EVE provides a flexible foundation for Industrial and Enterprise IoT edge deployments with choice of hardware, applications and clouds.
Top Use Cases› Consolidating a mix of container and VM-based workloads (e.g. for legacy apps) on
the same IoT edge hardware› Deploying edge hardware to serve as secure network proxy for downstream IoT
nodes and systems› Deploying out-of-band security and analytics apps leveraging a network SPAN port
Technical Summary› Fully-featured bare metal orchestration foundation› Targeted at the IoT edge: x86/Arm nodes with 1GB+ memory up to small clusters› Supports VM, OCI/Docker and Unikernel app deployment models › Supports zero trust security with all key functions built on HW root of trust› Enables zero touch onboarding with no device username/password required› Supports rollback/forward updates › Enables IO port disablement, CPU/GPU assignment to apps, distributed firewall› Open orchestration API for use with console of choice
› Extract data for local analysis / cloud and connect to new sensors
› Consolidate legacy and cloud-native workloads with no interference to existing setup
› Secure apps with private networks
IoT Edge Workload Consolidation
IoT Edge Router
Edge Security and Analytics
› Added security for current/legacy IoT installations
› Deploy a network proxy application (e.g., MQTT)
› Add app to update firmware of legacy hardware
› Deploy and manage security and analytics apps
› SPAN port collector on network enables non-intrusive, out of band data collection
› Gain visibility, monitor traffic and trigger events
LegacyHardware
IT, ERP, MES
IoT Edge Compute
EVE
Node
Node
Storage Service
WAN/Internet
Node
IoT Edge Compute
Network Probe
IoT Edge Compute
IDS
WAN/Internet
SPAN
PORT
Node
Node
Node
EVC EVC EVC
EVE EVE
Project Introduction: Fledge
120
Fledge is an open source framework and community for the Industrial Edge. Architected for rapid integration of any IIoT device, sensor or machine all using a common set of application, management and security REST APIs with existing industrial "brown field" systems and clouds.
Top Use Cases› Eliminate route based monitoring and deploy modern condition based
monitoring, predictive maintenance and situation awareness in industrial plants, mines and factories.
› Integrate IIoT with existing OT systems (no data silos)› Edge based anomaly detection, machine learning and AI to determine
machine state and/or part quality.
Technical Summary› Pluggable microservices based architecture to rapidly connect any new
or legacy machine, sensor or PLC using Python or C. › Easily build REST based applications and services that aggregate, buffer,
transform, analyze and deliver machine data from sensors to any/all OT systems and clouds.
› Consistent IIoT management and security APIs to scale up and out
121
Project Summary: FledgeStage 2- Growth
Project Release Status (as of Q12020)› Deployed in manufacturing, energy, water/waste water, and
oil and gas operations since Q1 2019.› Release 1.8 March 30, 2020
› ML/AI Tensorflow Support› Google Cloud North› Vibration Data Management 8000khz› New mgt/security API
• Monitors chemical levels in totes• Replace manual processes – RBM• KPI data for plant efficiency• Integrated w/ SCADA data
125
Fledge In EnergyPredictive Maintenance/Monitoring - Transformers
Series 4000Docker IoX Container
System ManagementCommercial Support
T&D Management IIOT Pub-Sub Engine
ERPTrouble Ticketing
B100 Transformer Monitor
• Data Collection & Aggregation• Edge Analytics• Alerting• IT-OT System Integration
• Monitors oil pump and fans• Monitors oil and air temp• Predicts transformer life-time• Eliminates break fix maintenance• Serves maintenance processes
126
Fledge In EnergyCondition Based Monitoring - Transformers
Series 4000Docker IoX Container
System ManagementCommercial Support T&D Management IIOT Pub-Sub Engine
ERPTrouble Ticketing
• Data Collection & Aggregation• Edge Analytics• Alerting• IT-OT System Integration
FLIR A310High-Low-Avg TempPer Object in Substation
Ethernet to Cisco 4000
127
Fledge In Municipal Water Condition Based Monitoring - Pumps
System ManagementCommercial Support
T&D Management IIOT Pub-Sub Engine
ERPTrouble Ticketing
• Data Collection & Aggregation• Edge Analytics• Alerting• IT-OT System Integration
FLIR AX8High-Low-Avg TempFocused on pump’s bearings
Ethernet to Nexcom
128
Fledge In MiningCondition Based Monitoring – Slurry Pumps
System ManagementCommercial Support T&D Operations & Management
• Data Collection & Aggregation• Edge Analytics• Alerting• IT-OT System Integration
Advantech1124
Current SensorHO21-100A 5B
Vibration SensorsX-Y-Z Planes
ADVANTECH ADC
Slurry PumpIngress/Egress Processing Plant
StorageBig Data Analytics
129
Fledge In ManufacturingSituation Awareness – Aircraft Paint Booths
System ManagementCommercial Support
T&D Operations & Management
• Data Collection & Aggregation• Edge Analytics• Alerting• IT-OT System Integration
AdvantechARC 1124
• Monitors paint booth micro-climates• Go/No-Go start paint process• Integrated with CNC machine and
autoclave status
ADVANTECH ADC
Dwyer TTE-109-W-LCDHHT-EU
Synergy and Engagement withOther LF Edge Projects› Fledge is working closely with Project EVE. Project EVE provides system and
orchestration services and a container runtime for Fledge applications and services. Together industrial operators can build, manage, secure and support all their non-SCADA, non-DCS connected machines, IIoT and sensors as they scale.
› Fledge is also driving towards opportunities with Akraino, with the Project’s verticals starting to roll out 5G and private LTE networks. Using Akraino blueprints, Fledge applications and services can be consistently managed as they utilize 5G and private LTE networks. Fledge’s first public use case was a private LTE network serving a wind farm where Fledge predicted maintenance conditions for turbine’s bearings.
130
Project Introduction: Home Edge Project
Concentrates on driving and enabling a robust, reliable and intelligent home edge computing open source framework and ecosystem running on a variety of devices at home. To accelerate the deployment of the edge computing services ecosystem successfully, the Home Edge Project provides users with an interoperable, flexible, and scalable edge computing services platform with a set of APIs that can also run with libraries and runtimes..
Top Use Cases› Service offloading in a home environment when device doesn’t have
required capabilities› Distributed computing framework to maintain low latency and high
data privacy
Technical Summary› Device/Service Management : Device/Services which are in home
network are discovered and managed based on their service capabilities.
› Service Offloading : Low end devices request high end devices to perform computing on behalf of them using the scoring manager.
› Scoring Manager : Helps to pick a right device to perform service
Home Edge: Drivers & EnablersDrivers1. Smart Home Products are now mainstream & need
common API/Gateway/UI/Lifecycle
2. AI technologies enabling learning and lifestyle/safety prediction requires local but connected Edge computing
3. Real time/low latency requirements increasing as safety, natural disasters and home health become mainstream beyond Telecom “triple play”
4. Data Storage & Data Privacy increasingly important and require sensitive data closer to home/userHome
Edge
Smart Home has a great potential to enable new business apps through home edge computing
Home Edge Executive SummaryHome Edge is a project targeted to› Enable Home Edge Computing Framework, platform and ecosystem
Edge Compute Framework for Smart Homes
Home Edge Scope› Define use cases, architecture and technical
requirements
› Develop and maintain the features and APIs targeting Smart Home use cases and requirements in a manner of open source collaboration
› Upstream the core features back to the existing/upcoming LF edge open source projects.
› Connect with Blueprints on Smart Home & Akraino through testbed validation
Technical requirements
› Dynamic device/service discovery at “Home Edge”
› Quality of Service guarantee in various dynamic
conditions (eg devices On or Off)
› Distributed Machine Learning
› Multi-vendor Interoperability
› User Privacy
Home Edge: High Level Platform Architecture
Device
Control
Container Runtime (based on EdgeX Foundry) / Deep Neural Net Runtime (Future)
Platform
Component
Edge Service
(Initial Seed
code)
Security
Home Device Control
Edge Setup
Edge Orchestration* Data Storage*
AI
Speech
RecognitionVision Service
Other AI
Service
Device Control
ServiceData Service
Other Edge
Services…
Legend
Machine Learning (Future)
Device Management
Service Discovery
Service Deployment
Edge Discovery
Command
Core Data
Data Interface
Metadata
Model Partition Converter
Distributed worker scheduler
Controller
Installer
Cloud
Interface
. . . . . . . . . . . . . . . . .
Device
Discovery
Controller
Discovery
Controller Adapter
NN Model Interface
Deep Neural Network Framework ( eg. TensorFlow Lite
)
. . . . . . . .
Air
Conditioner Distributed Job Executer
New Home Edge Apps and Services based on APIs
* Samsung seed code for Home edge computing platform architecture is based on EdgeX Foundry that is able to provide real-time, locality, and user privacy for various use cases, initially focused on Orchestration & storage
› Edge Device/App. discovery
› Device resource monitoring and preliminary service offloading
› Called “Scoring Manager” based on formula basis
› Will be improved by adopting Machine Learning (based on usage patterns)
Edge orchestration : Edge device/service discovery and remote service execution among edge devices.
❖ Edge orchestration features : Basic features of Service Management (Lifecycle management), Monitoring.
❖ Service Offloading to other device➢ Sharing of resource information for all the edge devices (CPU/Memory/Netowrk/Context)➢ Selecting Edge Device for service execution based on the capability of the device.
* Presented in IoT Solutions World Congress in Barcelona (@LF Edge Booth, Oct. 29th ~ 30th)
Home Edge: Use Cases – Anomaly Detection
Home Edge: Use Cases – Anomaly Detection
* Presented in IoT Solutions World Congress in Barcelona (@LF Edge Booth, Oct. 29th ~ 30th)
CCTV Camera(Low End Computing device)
CCTV Camera Tampering
Laptop(Master Edge Computing device)
Mobile(Edge Computing device)
Classify
Select a device to execute AI model
Device Faulty
Mobile video alert
(anomaly clip is sent)
Alert via AI Speaker
Use nearby camera to get
intruder detailsMobile Alert
Offload service to raiseservice Ticket
Notify User
Home Edge Computing
Anomaly detected
State of the Edge is an open source research and publishing project with an explicit goal of producing original research on edge computing, without vendor bias. The State of the Edge seeks to accelerate the edge computing industry by developing free, shareable research that can be used by all. The SotE Project contains LF Edge’s Glossary and Landscape projects.
Principles› The edge is a location, not a thing;› There are lots of edges, but the edge we care about today is the
edge of the last mile network;› This edge has two sides: an infrastructure edge and a device edge;› Compute will exist on both sides, working in coordination with
the centralized cloud
145
Project Summary: State of the EdgeStage 2 - Growth
Project Release Status (as of Q2020)› Moved under LF Edge April 2020
Baetyl offers a general-purpose platform for edge computing that manipulates different types of hardware facilities and device capabilities into a standardized container runtime environment and API, enabling efficient management of application, service, and data flow through a remote console both on cloud and on prem.
Top Use Cases› Light, secure, and scalable edge applications
› On drone processing› AI/ML- Allows for processing at the edge, reducing latency
› Quality Inspection by AI via video images› Automated/Zero touch onboarding- remote management
Technical Summary› Works with x86, arm, MIPS and OS agnostic› Services to speed development
› Video Ingress Service, ML Inference as a Service› For unstable Networks- has local persistence
146
Project Summary: BaetylStage 1-At Large
Project Release Status (as of Q12020)› Baetyl as a Container- Mid 2020› Remote Management- APIServer: Certification, application
Open Horizon is a platform for managing the service software lifecycle of containerized workloads and related machine learning assets. It enables management of applications deployed to distributed hyperscale fleets of edge computing nodes and devices without requiring on-premise administrators.
Top Use Cases› Management of ML models and containerized workloads on
constrained devices
Technical Summary› Provides a policy-based mechanism to securely deliver containerized
workloads to edge compute nodes of varying sizes and capabilities and in various connected states.
› Fully autonomous agent runs on every edge device to enable orchestration and manage the lifecycle of your containers
› Autonomous Agreement Bots (agBots) monitor each edge node› Model Manager automatically syncs assets bi-directionally based on
policy
147
Project Summary: Open Horizon
Project Release Status (as of Q22020)› Launched Q2 2020 with LF Edge and incubated with LF
Edge’s EdgeX
Where do I start?
Key Areas to look into
› Foundations – Key attributes› Community, Participation & Center of Gravity
› Not all Open Source created equal (from Github to Community Driven)
› Projects – What to look for› Diversity (LFX Insights), Contribution & Value created, Alignment with Standards
› Personnel – From Network Admins to DevOps › New ways of working (Fundamentals of Open Source Development, DevOps process & Training
from Network Admin to Dev Ops Admin)
› Vendors/SI/Service Providers – Direction› From closed to partially open to fully open
› End to End Open Source integration
E2E 5G Super Blueprint
Cloud Native 5G Demo
SOFTWARE COMMUNITY
Private LTE/5G ICN Akraino
ONAP 5G Use Case
THIS NEW BLUEPRINT FOR END TO END Open Source 5G
End to End
Super 5G Blueprint
5G Super BlueprintOverall Roadmap, building on production ready projects
CalendarQ2’21
5G Core + MEC
Kubernetes + ONAP + Anuket + Magma + local
breakout to MEC
5G E2E + Slicing + MEC
5G Core demo + commercial 5G RAN
Calendar
Q3’21
5GC + Slicing + O-RAN + MEC
5G Core demo + O-RAN-SC + E2E network slicing
Calendar
Q4’21
Components
Value
Use Cases
Fully disaggregated open source 5G core stack with
5G 101For consumers, enterprises and CSPs, 5G offers significant, compelling advantages over 4G:
› Capacity: 5G is designed to support a 100x increase in traffic capacity.
› Bandwidth (speed): 5G delivers up to 20 Gbps peak data rates and 100+ Mbps average data rates.
› Scaling (agility): 5G network slicing enables a single physical network to be partitioned into multiple virtual networks, each optimized for the different needs of specific users.
› Latency: 5G delivers more instantaneous, real-time access via a 10x decrease in end-to-end latency down to 1ms.
› Capability: 5G not only elevates mobile broadband experiences, but also supports new services such as mission-critical communications and large-scale Internet of Things (IoT) deployments.
› The 3rd Generation Partnership Project (3GPP) is the organization responsible for 5G standards. Two releases are relevant to current 5G deployments, while one provides an indicator of future capabilities:
› Release 15 in 2019 included support for enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC) and massive machine-type communications (mMTC) in a single network. It also supported 28GHz millimeter-wave (mmWave) spectrum and multi-antenna technologies.
› Release 16 in 2020 added standards for connected cars, smart factories, private networks and public safety to meet the needs of more diverse industries.
› Release 17, due in 2022 following delays as a result of COVID-19, will bring features such as URLLC for industrial IoT (IIoT), integrated access and backhaul (IAB), radio access network slicing for the 5G New Radio (NR), NR sidelink as well as support for multi-SIM devices for both LTE and NR.
› Provide feedback through VSFG› Host and staff a community lab› Answer questions› Give a talk / training› Create a demo› Evangelize LFE and its projects