IOT292-L

The extended cloud: AWS powers

edge-to-cloud applications

© 2020, Amazon Web Services, Inc. or its affiliates. All rights reserved.
What is the edge?
Edge use Why is it important
cases for customers?
Running workloads at the edge is challenging

Law of
PHYSICS

AWS Snowball Edge

AWS Outposts Law of
ECONOMICS Cloud
AWS IoT
Greengrass

Law of
AWS RoboMaker AWS Snowcone THE LAND
AWS for the edge: Giving you centralized control
and decentralized execution
Broad set of cloud services Lower cost of development
Reduced latency and edge-specific capabilities with single programming model

Seamless and integrated – Develop, connect, deploy, manage, and secure

it all with the same tools, same APIs, and the same control plane
 Edge-to-cloud architecture
Microcontrollers Networked PoPs & Regions
& IoT devices environments 5G telco networks Rugged edge On premises Metro centers (AWS Cloud)

24
Regions
(+5 announced)

77
Availability Zones
(+15 announced)

220+
PoPs

210+
Edge locations
 Edge-to-cloud architecture
Microcontrollers Networked PoPs & Regions
& IoT devices environments 5G telco networks Rugged edge On-premises Metro centers (AWS Cloud)

FreeRTOS AWS IoT SiteWise AWS Wavelength AWS Snowball Edge AWS Outposts AWS Local Zones

AWS IoT Greengrass AWS RoboMaker AWS Snowcone AWS Storage Gateway 24
Regions
77
Availability
zones
IoT Amazon Kinesis
Video Streams 220+
PoPs
210+
Edge
Amazon SageMaker Neo locations
 Edge-to-cloud architecture
Microcontrollers Networked PoPs & Regions
& IoT devices environments 5G telco networks Rugged edge On-premises Metro centers (AWS Cloud)

FreeRTOS

AWS IoT Greengrass

IoT Amazon Kinesis

Video Streams

Amazon SageMaker Neo

IoT: FreeRTOS

Local Cloud OTA &

connectivity connectivity Security code signing

Communicate with AWS Easily collect data and take Secure device data Deploy security updates, bug
IoT Greengrass devices actions on microcontroller and connections fixes, and firmware updates
without a cloud connection -based devices to devices in the field

FreeRTOS is open source

 New feature

IoT: FreeRTOS Long-Term Support (LTS)

generally
available

Get predictability and feature stability for

two years with FreeRTOS LTS libraries

</>

Feature Security updates and Updateable, secure

stability critical fixes connectivity
IoT: AWS IoT Greengrass

Security and Local Local

Secrets Manager Messaging Actions

00011100
0
1 1 1
1
1 1
1 1
0 0
0 0
0 0
0 0

Container Support Stream Manager ML Inference

 New feature

IoT: AWS IoT Greengrass 2.0

generally
available

A flexible edge platform to easily build, deploy, and

manage device software across millions of devices

Completely Deploy device Locally deployable

modular software at scale and configurable

New! AWS IoT Greengrass is now open source!

IoT: AWS IoT Core

Identity Device Message Rules Device Registry

Service Gateway Broker Engine Shadow

Manage authorization Fully manage Reliable and fast Ingest large amounts Understand and Define and catalog
of devices and connectivity optimized communication of IoT data at a low control the status device for easy use by
provision unique for IoT workloads across your cost, pre-process it, of your device at AWS services
identities at scale IoT fleet and make it available any time
to 10+ services for
analytics, reporting,
and visualization
Solution AWS is how Outcome
Siemens MindSphere, industrial AWS-first platform provides Connect up to 80% of
IoT as a service, powers IoT secure asset onboarding to worldwide industrial
solutions from edge to cloud AWS IoT to collect and analyze automation and edge devices
with advanced analytics and AI, data from connected products,
built on AWS plants, and systems Enable high
uptime and SLAs
Standardized reporting
and asset monitoring
Machine learning: Amazon SageMaker Neo
Train once and run anywhere with 25x performance

SageMaker algorithms

ONNX ANDROID

iOS / MAC

LINUX
SageMaker Neo
WINDOWS

Broad framework support Broad operating system and

hardware support
 New feature

Machine learning:
generally
available

Amazon SageMaker Edge Manager

Operate machine learning (ML) models efficiently across fleets of smart devices
 New service

Machine learning: preview

AWS Lookout for Equipment

Detect abnormal machine behavior using existing industrial sensor data

Detect machine abnormalities in real time

Specifically developed for industrial assets such as turbines,

motors, pumps, and more

Automatically identifies equipment anomalies by learning

the healthy state and operational relationships between
sensors on each asset

Enable predictive maintenance with no ML expertise needed

 New service

Machine learning: Amazon Monitron

generally
available

End-to-end equipment condition monitoring system, powered

by machine learning, enabling predictive maintenance

Easy to install, and easy to use

Technicians can start monitoring equipment in less than an hour,
measuring temperature or vibration, compared to an
ML-determined or ISO standard baseline

No machine learning experience required

No development work or machine learning expertise is required

Continuously improving
Feedback helps to continuously improve machine learning model
performance while hardware is updated over-the-air (OTA).
 “Over the past year we worked with AWS to help develop
the critical, but sometimes overlooked part of running a
successful manufacturing business, knowing the
condition of your equipment. For manufacturers
worldwide, maintaining equipment uptime is the only
way to remain competitive in a global market.
Unplanned downtime is costly both in loss of production
and labor due to the firefighting nature of breakdowns.

The Amazon Monitron condition monitoring system can

give both large industry manufacturers, as well as small
Fender Musical Instruments ‘mom and pop shops,’ the ability to predict equipment
Corporation is an iconic failures, which gives us the opportunity to
brand and a leading
manufacturer of stringed
preemptively schedule repairs.”
instruments and amplifiers Bill Holmes
Global Director of Facilities
Fender
Analytics: Amazon Kinesis Video Streams
Amazon Kinesis Video Streams is a fully managed service that elastically scales to
support the media streaming needs of millions of camera-enabled IoT devices

Low-latency livestreaming Large-scale secure media Real-time and on-demand

and two-way interactivity ingestion and retention playback and processing

WebRTC audio, Encrypted, durable, and HLS, MPEG-DASH, and integration

video, and data time-indexed media storage with AWS machine learning services
About Solution Outcome
Wyze is a smart-home Amazon Kinesis Video Streams Scaled to seven million smart
technology company based in automatically provisions and cameras in less than 3 years
Kirkland, Washington. In 2017, elastically scales all the
Wyze launched its first product infrastructure needed to Reduced development time
– the Wyze Cam, an affordable ingest streaming video data for new features by 50%
smart-home camera. Today, from millions of devices
Wyze offers a growing system Securely stores video data
of smart-home devices from millions of devices
Enables continuous-motion
video recording and streaming
Facilitates security compliance
 New service

Machine learning: generally

available

AWS Lookout for Vision

Improved defect detection speed, consistency, and accuracy

compared to manual inspection

Lower cost of quality control inspection compared to

machine vision systems

No ML experience required – start building a model with as

few as 30 images
 New service

Machine learning: preview

AWS Panorama Appliance

Build and manage computer vision applications at the edge

Turn existing onsite cameras into powerful edge devices

Analyze video feeds from multiple cameras in parallel, and

generate highly accurate CV predictions within milliseconds

Bring your own or select partner CV applications such as PPE

detection, social distance monitoring, license plate detection,
and more

Get started with the AWS Panorama Appliance Developer Kit

 New service

Machine learning: preview

AWS Panorama Device SDK

Build new cameras that enable computer vision at the edge

Build new edge gateways and smart cameras with

AWS Panorama using the AWS Panorama Device Software
Development Kit (SDK)

Includes a device software stack for computer vision,

SDK sample code, APIs, and tools to test Panorama device
software on hardware

Supports leading silicon solutions: NVIDIA Jetson product

family and Ambarella CV 2x product line
 “Computer vision helps us innovate and optimize
several processes, and the applications are
endless. We want to use computer vision to assess
the size of trucks coming to our granaries to
determine the optimal loading dock for each truck.
Cargill brings food and We also want to use computer vision to understand
agricultural, financial, the movement of assets in our plants to remove
and industrial products to bottlenecks.
people who need them,
all around the world AWS Panorama enables all of these solutions with
a managed service and edge appliance for
deploying and managing a variety of local
computer vision applications.”
Victor Caldas
Computer Vision Capability Lead
Cargill
 Edge-to-cloud architecture
Microcontrollers Networked PoPs & Regions
& IoT devices environments 5G telco networks Rugged edge On-premises Metro centers (AWS Cloud)

AWS IoT SiteWise

AWS RoboMaker
IoT: AWS IoT SiteWise
AWS IoT SiteWise collects data from the plant floor with a local gateway,
structures and labels that data, and generates real-time KPIs and metrics
to make better data-driven decisions

ASSET MODEL
1010110
01010 10
0101001 Property: Attribute
01010 10
1010110
1010101
01010 10 tool_id :
0101001
01010 10
1010101 Property: Time-series
101010101010101
010101010101010 101011010
101010101010101 01010 10101 pressforce-ch1 :
0 1 0 1 0 1 0 1 0 1 0 1 0 1 00 1 0 1 0 0 1 0 1 0
1010
01010 10101 Property: Formula
1010101010
1010101010 Avg Pressforce : f
1010101010
(pressforce-ch1)

Ingest equipment data Structure data and specify Store data in a time-series- Create and share dashboards
into AWS in minutes performance metrics for your optimized data store to visualize live and historical
equipment and processes equipment data
Mission Solution Impact
Volkswagen Group is Volkswagen and AWS are For their component production
transforming its business to developing the Volkswagen processes, Volkswagen can
become the global leading Industrial Cloud, which includes reduce administrative efforts
provider of sustainable mobility foundational platform services, through automated data
retrieval and reporting, achieve
and to improve production spanning the edge to the cloud sustainable improvement of
efficiency by 30% to enable various use cases machine availability through
transparent activity tracking and
To achieve that, Volkswagen The manufacturing shop floor knowledge sharing across
needs a flexible, comprehensive data is ingested through AWS plants, and increase
and standardized industrial IoT IoT SiteWise, stored in a data productivity by having full
platform that can ingest and lake, and used by a custom web visibility into production losses
combine data from all of its application to monitor near- and their influence factors
machines, plants, and legacy real-time status of machines
applications and calculate overall equipment
effectiveness (OEE) for the
cylinder production line
Robotics: AWS RoboMaker

Build Test & train Deploy

Connect robots to AWS services Fully managed cloud simulation service Update robots in the field
for machine learning, storage, with AWS IoT Greengrass
video, speech, and more with ROS Automated 3D world generation
Cloud Extensions
Need Challenges Solution
Increase code release velocity while Access to hospital environment was Automated simulation-based
improving test coverage and safety difficult, so testing was costly and regression testing in a CI/CD
for running robots in hospitals time-consuming pipeline using AWS RoboMaker
and the hospital simulation
Test coverage was limited and discovery environment
of software defects was delayed

“AWS RoboMaker simulation helped us save money and time – and eliminate obstacles to scaling our
testing efforts. With RoboMaker, we can forget about infrastructure and focus on development and
testing. We implemented a RoboMaker simulation CI and can now easily test our robots in different
environments, resulting in a huge quality improvement.”
Goker Ertunc
Co-Founder & CIO, Milvus Robotics
You need to know where you are at the edge

Delivery Asset tracking Visualization Customer engagement

 New service

Introducing: Amazon Location Service preview

Securely and easily add location capabilities to your applications

Maps
Easy access to high-quality data

Cost-effective
Tracking Places

Your Built-in privacy and security

application

Shorter time to production

Geofencing Routing
(Coming)
 Edge-to-cloud architecture
Microcontrollers Networked PoPs & Regions
& IoT devices environments 5G telco networks Rugged edge On premises Metro centers (AWS Cloud)

AWS Wavelength
Infrastructure solutions: AWS Wavelength
AWS compute and storage services within the telecommunications
providers’ data centers at the edge of 5G networks

Low latency / Consistent Same AWS benefits Broad coverage

high bandwidth development
experience
Edge data processing and inference
“With critical GPU workloads rendered
from AWS Wavelength coupled with 5G’s
bandwidth, we can achieve the low latency
and increased connectivity to deliver a
pathway to truly connected factories,
where data flows seamlessly between
machines and cloud to enable AI-powered
real-time decisioning at the edge.”
Shanky Viswanathan
CTO Communication,
Media and Information Service Business Group,
Tata Consultancy Services
 Edge-to-cloud architecture
Microcontrollers Networked PoPs & Regions
& IoT devices environments 5G telco networks Rugged edge On premises Metro centers (AWS Cloud)

AWS Snowball Edge

AWS Snowcone
AWS Snow Family: The evolution

Data movement Edge computing

Customer sites Edge sites

• Set of services and functionality on the partially to fully disconnected and

secure edge
• Enables data collection, data processing, and data movement from the edge
to AWS
AWS Snow Family: The rugged edge
Challenge Solution Outcome
Novetta builds and deploys Edge processing and application Range performance exceeded
advanced analytics solutions server run on an AWS Snowball Edge expectations; team set up the
at the edge for public sector compute-optimized device, enabling entire stack in under 20 minutes
defense customers that Novetta to bring data and Amazon
improve situational awareness EC2 instances to the field and Mapped vessels and personnel
during crisis / disaster operate disconnected from the cloud from 30+ organizations on the
response operations same map covering 35 sq miles
of open water
 Edge-to-cloud architecture
Microcontrollers Networked PoPs & Regions
& IoT devices environments 5G telco networks Rugged edge On premises Metro centers (AWS Cloud)

AWS Outposts

AWS Storage Gateway

Infrastructure solutions: AWS Outposts
Extend AWS infrastructure and services on premises

AWS-designed, Nitro infrastructure for

the same security and performance
Automatically managed and updated as part
of AWS Regions
Single point of service and support
Build once, deploy anywhere

Customers want the same experience

on premises and in the cloud
 New service

AWS Outposts in 2 new sizes pre-announcement

Run AWS Outposts in locations with limited space or capacity requirements

Offers the same AWS infrastructure, services, APIs, and

OUTPOSTS
tools on-premises, now with a smaller form factor

Choose between a 1U Outpost server with an AWS Graviton2

processor or a 2U Outposts server with an Intel processor
1 ¾ inches tall

Run AWS services locally, including EC2, ECS, and EKS

and edge services like AWS IoT Greengrass

Ideal for workloads that require low latency and local

3 ½ inches tall
processing needs

80 inches tall
Simple device installation by either your own
on-premises personnel or a preferred third-party vendor

Amazon Confidential
On-premises storage: AWS Storage Gateway
Provide on-premises access to virtually unlimited cloud storage

Continuous
Migration Modernization reinvention

Move backups to Use on-premises file shares Low-latency access for

the cloud backed by cloud storage on-premises applications
to cloud data

For all stages of your cloud adoption journey

Challenge Solution Outcome
Multi-year cloud migration journey AWS Outposts on premises Lift and shift: Build once, run on
with goal of being all in on AWS running VMs and premises, and seamlessly migrate to
containerized workloads AWS Region when ready
Provide low-latency, real-time
investment research and insights Truly consistent hybrid Accelerate adoption of AWS across
experience with fully development teams
Eliminate overhead associated with managed and monitored
managing and monitoring different AWS services both on Improved business operations and
platforms for on-premises and premises and in the cloud end user satisfaction
cloud workloads

“As we continue to modernize our data collection and software, we want to create a containerized hybrid
infrastructure and easily deploy using AWS services. With AWS Outposts, we can build once and run applications
on premises and easily migrate our applications to an AWS Region where possible. Ultimately, this allows us to
accelerate the adoption of cloud technologies within our development teams, keep up with accelerating business
and customer needs, and support our long-term journey to the cloud.”
James Rhodes, Chief Technology Officer, Morningstar
 Edge-to-cloud architecture
Microcontrollers Networked PoPs & Regions
& IoT devices environments 5G telco networks Rugged edge On premises Metro centers (AWS Cloud)

AWS Local Zones

Infrastructure solutions: AWS Local Zones
Run applications that deliver ultra-low latency to end users

AWS infrastructure Local compute, Connect to services Deliver new

at the edge storage, database, in AWS Regions low-latency
and other services applications

Support for Amazon EC2, Amazon ECS, Amazon EKS, Amazon VPC, and Amazon EBS
Centralized control with decentralized execution

Connect Deploy Manage Secure

• Stream Manager • AWS IoT

• Iridium Connect Greengrass 2.0
AWS IoT AWS IoT • Container support AWS Amazon AWS Key
Greengrass Greengrass CloudFormation SageMaker Management Service

• AWS IoT Core

for LoRaWAN
AWS AWS Amazon AWS AWS Secrets
IoT Core Lambda CloudWatch IoT Core Manager

AWS AWS IoT AWS Config AWS IoT Device AWS IoT
Wavelength SiteWise Management Device Defender

• Fleet Hub
• Fleet indexing
AWS IoT Device AWS Systems
Management Manager
We have experience doing
this for ourselves
• Globally, we have more than
175 fulfillment centers and
more than 40 sort centers
• We have opened more than 50 robotic
fulfillment centers around the world
• Amazon currently uses the help of
more than 250,000 robotic drive
units around the world
• Amazon has over 20,000 trailers
on the road
• Customers have purchased more than one
hundred million Alexa-enabled devices
Pulling it all together
IoT
Robotics
Real-time video
Machine learning
Edge computing
Predictive maintenance and operations
Putting it all together: Woodside LNG Plant
More edge-specific capabilities,
more services, and
more features within those services,
than any other cloud provider
Explore re:Invent sessions for a deeper dive into
each layer of the edge-to-cloud architecture
Session ID Session Title
CMP208 AWS Outposts: Cloud where you want it
IOT291-L Connect today, transform tomorrow with AWS IoT
STG219 Edge computing innovation with AWS Snowcone and AWS Snowball Edge
ZCW210 Building robots to help hospitals become safer and smarter
TLC304 At the cutting edge: Dedicated edge networks and mobile edge cloud
IOT303 Developing and deploying modern edge applications at scale
CMP311 AWS Wavelength: Run apps with ultra-low latency at 5G edge
IOT207 AWS at the edge: Using AWS IoT to optimize Amazon wind farms
AES19 Connecting the world of tomorrow, today, with 5G edge cloud
WPS208 Drones and Snowballs: Delivering imagery at the edge
IOT314 Managing industrial data at the edge with AWS IoT SiteWise
Thank you!
Bill Vass
Vice President of Engineering

© 2020, Amazon Web Services, Inc. or its affiliates. All rights reserved.