4. • Global Technology Hub
• Years of semiconductor/foundry experiences
• Electronic component/product design experiences (OEM/ODM)
• Lab Engineers with chip/firmware expertise
• Regional Advantages
• Proximity to the larger region of manufacturing hub
• Government support for hi-tech manufacturing
Why IoT Lab in Taiwan – Our Goal?
5. • Build local solutions
• Getting devices qualified
• Getting solution design and fine tuned
• Deploy global products
• Hardware from OEM/ODM
• Applications & solutions from partners/IoT Lab
• Powered by AWS platform
Why working with IoT Lab – Our Vision?
6. IoT Platform for Growth
Problem
LG’s new smart products include embedded WiFi chips
for communication and AI technology for learning
about their user’s behavioral patterns. LG built its own
IoT Platform but experienced difficulties in supporting
growing number of connected devices.
Solution
Migrated its platform of 1000+ servers from its data
center to AWS. Using serverless architecture with AWS
Lambda to register and mange devices on the cloud
through IoT Core, and store data in Amazon ElastiCach
and DynamoDB.
Impact
By using AWS for the new IoT Platform, LG saved 80
percent in development cost.
We found that it was better to receive help from
AWS solutions architects compared to solving
problems ourselves. We enables developers to focus
on writing business logic for LG service scenarios.
Kim Kunwoo
Chief of the Service Development Team
LG Cloud Center, LG Electronics
“
”
7. • Business Opportunities
• Connect hardware partners with AWS customers
• AWS credits and benefits
• Marketing Opportunities
• AWS Partner Device Catalog
• Branding and co-marketing opportunities
• Technical Differentiation
• Integration with AWS platform
• Expertise at AWS IoT Lab
Why Device Qualification Program?
8. Partner Device Catalog
As customers evolve in their digital transformation
journey, they are looking for ways to quickly and
efficiently move certain workloads to the edge.
Collaborating with AWS helps us bring the best of
the cloud to the edge on Lenovo infrastructure and
devices. Lenovo has qualified industry-leading
servers, PCs, and intelligent devices through AWS’s
simple and easy to use AWS IoT Device Tester
qualification tool found in the AWS Device
Qualification Program for AWS IoT Greengrass and
listing the hardware in the AWS Partner Device
Catalog.
Wilfredo Sotolongo
VP and GM of IoT
Lenovo Data Center Group
“
”
9. 2017 2018 2019
Greengrass GA
Jun 2017
FreeRTOS GA
Nov 2017
87 Greengrass Devices
34 IoT Core Devices
50 KVS Devices
Mar 2019
74 Greengrass Devices
32 IoT Core Devices
re:Invent 2018
50 Devices
Sep 2018
AWS IoT Lab
Started
Oct 2018
5 Partner Kits
Sep 2018
10 Partner Kits
re:Invent 2018
12 Partner Kits
Mar 2019
Device Qualification Program and
Device Tester Development
(Greengrass+ FreeRTOS)
Aug-Oct 2018
AWS IoT Lab – Timeline
Alexa/IoT POC
IoT/ML/Summerian POC
IOT/BI POC
Nov 2018
AWS IoT Lab TW
Announced
Mar 2019
Alexa/IoT/ML POC
Feb 2019
Greengrass
10. ERP
PLM
APS
Prototyping, Pilot Production, Machine
Evaluation, System Integration and
Scale-up
Large scale demonstration site,
integrated test environment of Industry
4.0
VMX = Virtual, Multi-machine (different
brand of manufacturing machinery)
Integrated with AWS (Edge Computing +
IoT Gateway + Cloud)
Intelligent Machinery Technology
Innovating a better future
17. Vortex Data River enabling vendor-
neutral device ecosystem
Integrated with AWS IoT Greengrass
connecting with AWS IoT platform
Complete, efficient data acquisition
leads to way to the business of the
future
Edge Computing Platform
Enabling Data-to-Decision
18. ADLINK
“AIoT at the EDGE”
Chia-Wei Yang
Business Development Manager
ADLINK Technology
19. ADLINK Overview
Over 20 Years of Embedded Experience
>1900
Employees
Established
August
Publicly Traded Since Taiwan Stock
Exchange Listing
USA
Canada
France
India
Korea
Israel
Chairman
Jim Liu
President
Daniel Yang
Revenue (USD)
Capital (USD)
$73M
Headquarters in
Germany
UK
Singapore
China
Japan
●
●
●
●
1995
Taipei,
Taiwan
2002 TAIEX: 6166
$297M (Y2016)
$350M (Y2017)
Market Cap (USD)
$468M
●
20. AIoT at the Edge
Data 2 Decision
Convert Info to Decision & Action to enhance overall
Efficiency
Data River
Steam the RIGHT INFO to RIGHT Place @ RIGHT Time
Heterogeneous Computing Platforms
Powerful EDGE, Integrated CPU, VPU, FPGA, GPU
Industrial
Automation
Telecom &
Networking
Test &
Measurement TransportationMedical /
Healthcare
Infotainment &
Vending
Military &
Aerospace
AWS
21. Source: en.ippon.tech
Engineering Collaboration
Accelerating IoT - EDGE Development
• Collaborating in certificating
AWS IoT Greengrass with
ADLINK EDGE Devices
• Enabling 3rd party DL / ML
modeling builds on AWS
• Integration of AWS RobotMaker
Service and ADLINK ROS2.0 Robotic
Platform
22. Synergy for Factory of the Future
AI Machine
Vision
Industrial
IoT
Robotic
(Mobile)
24. Connectivity to AWS IoT platform
Accelerated ML inference with Intel
Movidius and OpenVINO
Ubiquitous solution designed for each
application domain from industrial IoT
to smart retails.
Industrial IPC Solutions
Designed for each industrial and business
applications
35. Integrating hardware devices, software
applications, and the cloud for
customers.
Innovative solutions adapt to different
vertical industries with device and
cloud working in harmony.
Integrate protocol by click and done
(BLE, Z-Wave, CAB Bus, ZigBee,
ModbusRTU, Modus TCP, WiFi)
Make IoT A Reality
One day, all devices will be connected to the
Internet and will be given intelligence
43. Target
• Smart retail analytics and video/image analytics
partners
• Smart retail facility operators
Problem
• Need actionable data to engage customers,
optimizing the business and making smarter data-
driven decisions.
• Need accurate, real-time insights of the audience’s
spontaneous behavior, interest and demographic
profile.
POC/Solution
• Based on sophisticated Computer Vision and Face
Analysis algorithms, the software enables detection
and measurement of multiple faces in a video
frame. The software used in combination with a
HD/UHD4K camera sensor to detect and track
audience facial information.
• Actionable data (Viewers/OTS, Dwell Time,
Attention) with anonymous audience visual analysis
with customer ML model (Demographics, Age,
Gender, Location and Mood Insights)
IoT Core
Amazon
QuickSight
Amazon
S3
Kiosk
w/ Video Analytics
Amazon
EMR
(w/ Zeppelin)
Smart Retails Prototype
44. Amazon
Lambda
Amazon
SNS
Amazon
SQS
IoT Core
Target
• Smart home device manufacturers
• Smart facility solution partners
• Smart school
Problem
• Hardware manufacturer have limited knowledge
with Alexa/IoT integration
• Challenges with best practice on integrating the IoT
platform with Cloud
POC/Solution
• Prototype GreenGrass/FreeRTOS/IoT SDK plus
Alexa AVS/AIS API
• Prototype best practice OTA, Device Management,
and serverless lambda functions.
• Include customer facing application template for
device registration/provision.
• Enable Alexa integration smart home solution using
AWS IoT platform (and documented for other smart
home product manufacturers).
Smart Home Prototype
45. Alexa and IoT Integration
Problem
Looks to improve customer loyalty by delivering new
features for Vizio’s customers such as voice control.
However, Vizio don’t want to build an IoT solution from
scratch to deliver this improved customer experience.
Solution
Working with AWS, we created a secure and scalable
implementation using AWS IoT platform and serverless
architecture to deliver Alexa voice skill to our TVs.
Impact
Delivered a working prototype in a few days, and soon
after delivered new Alexa voice skill to millions of TVs in
the field. Vizio now have a high scalable IoT backend
build on AWS, which can now power many more
applications that will deliver new features.
Working with AWS has been one of the most
positive engagement with a technology company.
Bill Baxter
Chief Technology Officer
VIZIO
“
”
46. Application
(SNS, SQS, APIG, Cloud9)
Target
• Smart retail hardware/solution partners
• Smart Facility/Building/Hotel/Community operators
Problem
• Lack of end-to-end solution integrating IoT/AI
platform with Operational Tech
• Lack of resources for vertical application and a
marketplace for these applications
POC/Solution
• Work with partners and customers to create
technology prototypes and templates for:
• Facial/People Recognition (VIP, Intrusion Detection, guest
registration, guest access permission)
• People Tracking (tailgating, people search)
• Operational tools, provisioning tools, and management tools.
• Analytics (traffic patterns, hot spot for guest visit)
IoT Core
Smart Store Front
Data Lake
(S3, Dynamo, Redshift)
Analytics
(Athena, Kinesis, QuickSight)
AI
(SageMaker, Rekogition)
MQTT
(Meta Data/Shadow Update)
Direct Secure Connect
(Raw Data, Log, Photo, Video)
Smart Facility/Hotel POC
47. Application
(SNS, SQS, APIG, Cloud9)
Target
• Smart device provisioning/manufacturing partners
• Smart utilities operators, community owner
associations, city planning authority and data carrier
Problem
• Lack of end-to-end solution integrating IoT and
provisioning platform with operational tech
• Lack of market-ready templates for big-data/ML
analytics
POC/Solution
• Work with partners and customers to create
technology prototypes and templates for:
• Facial/people Recognition (VIP, people search, intrusion
detection).
• Object Recognition (license plate, parking, toll).
• People/Object Tracking (tailgating, Left passenger,
unattended object/bag/trolley, parking clearance).
• Operational tools, provisioning tools, and management tools.
• Analytics (traffic patterns, hot spot).
IoT Core
Smart Sensors/Meters
Data Lake
(S3, Dynamo, Redshift)
Analytics
(Athena, Kinesis, QuickSight)
AI
(SageMaker, Rekogition)
MQTT
(Meta Data/Shadow Update)
Direct Secure Connect
(Raw Data, Log, Photo, Video)
Smart City POC
Field applications for
provisioning and maintenance
48. Trusted IoT Platform
Efficient, cost effective serverless platform
Support most popular ML frameworks
Why working with AWS on AIoT?