Digital Transformation is accelerating as manufacturers double down on onshoring and near-sourcing while adding KPI’s focused on resilience and adaptability, strengthening the ROI of digital transformation projects. We’re now well into OT/IT convergence with IT technologies permeating OT, reducing costs and increasing the speed and scalability of deployment. This session will explore the leading technologies moving to the industrial edge of OT. Driving virtualization technologies to the edge of networks is a key enabling technology that will soon become one of the keys to success in digital transformation.
6. Building Solutions on Open Architectures
• Reduces vendor lock-in
• Opens massive ecosystem of hardware and software
providers
• Maximizes utilization of compute resources, reducing
required hardware footprint
• Scalable for growth and flexible to adapt to yet-
undiscovered use cases
7. • Using a Hypervisor, one piece of
hardware can run multiple “virtual”
machines
• Each VM bundles an OS, the application
and any of it’s dependencies
• VM’s are easily replicated across
different hardware platforms
In the world of data centers, we
virtualize with Virtual Machines
(VM’s)
Virtualization as a tool for scale: Virtual Machines
Hardware
Hypervisor
App 1 App 2 App N
Guest
Operating
System
Guest
Operating
System
Guest
Operating
System
Virtual Machine Virtual Machine Virtual Machine
8. • Because we replicate the OS with every
VM, they have a lot of overhead
• When we move out of the data center
towards ”the edge” we require hardened
devices, fanless, wider temp ranges,
smaller physical footprints
• Conclusion: VM’s are a better fit in data
centers than “the edge”
But can we take VM’s out of the data center?
Hardware
Hypervisor
App 1 App 2 App N
Guest
Operating
System
Guest
Operating
System
Guest
Operating
System
Virtual Machine Virtual Machine Virtual Machine
9. • Containers provide a lower-overhead
virtualization solution
• Containers share the Host OS so are
much leaner than VM’s
• Each container holds the application
and its dependencies, libraries and
settings, decoupling it from the physical
hardware.
• Containers are isolated from each other
and the outside world , interconnections
can be made over a virtual network
A leaner form of Virtualization: Containers
Hardware
Docker
App
1
Container
Host Operating System
Libs
App
2
Libs
App
3
Libs
App
…
App
N
Libs
Container Container Container Container
10. 1. Resilience
a) Each container is isolated from the next. One crashed app doesn’t bring down the
machine
b) Each container can be configured to auto-restart if there is a crash
Advantages of Virtualization via Containers
11. 2. Portability
a) Containers contain both the app and any dependencies. This allows you to migrate
easily across hardware platforms, including ARM and x86.
Advantages of Virtualization via Containers
12. 3. Scale
a) New containers can be deployed in seconds making it easy to deploy one solution
thousands of times, even across varying hardware platforms. No more, “It worked
on my machine.”
b) Because it’s portable, solutions can be deployed on larger or smaller hardware
footprints.
c) Because we haven’t created one massive monolithic application, we can easily
move containers up or down the stack.
Advantages of Virtualization via Containers
13. 4. Security
a) Each container gets its own virtual network stack with no access to sockets or
another container other than connections specifically configured for that container.
Managed correctly, this can significantly reduce the number of attack vectors
b) Note that ensuring security features are properly configured remains the
responsibility of the solution integrator.
Advantages of Virtualization via Containers
14. 5. Speed and agility
a) Need to update one of the virtualized applications? Containers are easily stopped,
started, updated.
b) Want one of your applications to always use the most recent version? Or want it to
always use a fixed version?
c) Discover a new use case that requires adding another application?
d) Is your solution growing wildly in scale? Need to share or balance resources across
hardware platforms? Add a container orchestration tool like Kubernetes to the mix.
Advantages of Virtualization via Containers
15. • Docker is the dominant flavor of container
technology with massive adoption since 2014
• Docker is part of the Linux Foundations Open
Container Project, so even other container
technology supports Docker images
• From the market, we’re beginning to see the
ability to run Docker containers as a
requirement in hardware RFP’s as customers
embrace open architectures and containerized
solutions
Container Technologies: Docker
16. Access Edge
On-prem Data Center Edge
Smart Device Edge
Constrained Device Edge
Server based compute in secure
locations
Microcontroller based edge devices
IoT (headless) and end user client
compute
Server based compute Edge
Exchange
Server based compute Internet
Exchange (IX)
Centralized Data Center
SERVICE
PROVIDER
EDGE
USER
EDGE
Server based compute in traditional
cloud data centersØ Using containers enables “cloud native”
architectures to be extended to the Smart
Device Edge.
Ø Each component of the solution (database,
logic engine, visualization, etc) is an
independent service
Ø This enables components of the solution to
easily migrate up the stack as applications
grow.
Ø “Cloud native” architectures aren’t just for the
cloud!
Virtualization extends Cloud-Native Architectures
to the “Edge”
Regional Edge
Edge Taxonomy: Linux Foundation Edge
17. Use Case:
A multinational manufacturing
company wants to monitor uptime of
their machines.
Example - Scaling Container based Virtualization
How Cloud Native Architectures and Containers simplify and
accelerate IIoT solutions development
18. • The PoC scope is to monitor a manufacturing cell of 5
machines with Modbus TCP communications
• The application is to provide, alarms and reports, and
eventually machine learning inference capability.
• UNO Industrial PC with Ubuntu, Docker, Ignition,
InfluxDB, PostgreSQL, Node-Red, AI engine.
Phase 1: Smart Device Edge Implemented PoC
Hardware: UNO Industrial PC
Docker
Ignition
Container
Ubuntu Linux
Libs
AIinference
engine
Libs
NodeRED
InfluxDB
PostgreSQL
Libs
Container Container Container Container
LibsLibs
19. • The boss loves it! Now, how long will it take you to
expand the PoC to a plant-wide solution for 100
manufacturing cells?
• We keep the architecture and components, but move
some containerized services.
• Lets deploy Ignition and a database to the On-prem
Data Center Edge, keeping some data, processing and
inference capability at the Smart Device Edge
Phase 2: Expand to Entire Plant
20. • The CEO wants it company-wide! Lets scale to 6000 manufacturing cells over 80 sites in 6 countries.
• Again, keep the architecture, but lets make some more moves.
• Let’s move Ignition and add an instance of Influx to a cloud based datacenter.
• Let’s also add NodeRED on the On-Prem datacenter to use as a logic engine, publishing just the outcome
data up to the cloud, keeping the detailed raw data historian in the on-prem datacenters to keep costs low.
Phase 3: Let’s go Global
21. • Virtualization and containers can enable large scale deployments with reduced
engineering cycles
• When you start to scale, its critical to address items that can quickly become pain-
points;
• zero-touch provisioning
• zero-trust security
• device management
• application management
• And more…
Considerations for scale
22. ØOpen architecture solutions provide access to a massive ecosystem, don’t re-
invent. Use the tools available to create your unique solution!
ØVirtual Machines work well in data centers, but are too heavy for the User Edge
ØUse Containers to create virtualized solutions that extend to the Smart Device
Edge
ØCloud-native architectures provide agility and scalability, with or without the cloud
ØThink about tools required to scale, before you scale!
Summary