>> Hi everyone, welcome
to IoT Show Deep Dive. My name's Pamela Cortez. I'm from the Azure IoT team and
I will be your host for today. Deep Dives are all about
showcasing an end-to-end solution, deeper dive against some scenario. So, I'm really excited
that this first Deep Dive, we're going to be
focusing on end-to-end industrial solutions
with PTC and Azure IoT. So, I have special guests here, Shafia Waheed from
the Microsoft Azure IoT team. I also have Joe Byron, he is the CTO of the IoT at PTC. I've made sure to
stay that correctly. >> You did a good job. >> So, we're going to go ahead. We're going to deep dive
into end-to-end solution. If you guys have any questions, I will be in the chat and
answering your questions and also asking Joe and Shafia
those questions as well. So, feel free to ask those
questions throughout the deep dive. All right. Let's go
ahead and get started. >> All right, thanks
Pamela. So, hi everyone, I'm Joe Byron from PTC. A little bit about
PTC and a bit about Industrial IoT which is what
we're really focused on here. PTC is a company that focuses on software solutions for product manufacturers and
industrial innovation. So, what we mean by that is, we're helping people
who create and use physical products and
machinery in industry, we're helping them to
optimize the creation of those products and the life cycle
management of those products. So, lately, we've been really
excited about Internet of Things and Augmented Reality technology as
it applies to those use cases. So, we thought we'd take
some time here today to get you a sense of what
Industrial IoT is all about and how we work
with Microsoft in building solutions that satisfy
these challenges in industry. >> Thanks Joe and welcome. Indeed, in the past
few months actually year we have been
working together with PTC to bring these
Industrial IoT solutions to our joint customers. Today, we will deep dive
into how PTC with the set of solutions are building IoT solutions for our customers on
top of Azure IoT. Of course, feel free
to ask your questions, we'll be happy to answer them. Let's get it started Joe. >> All right, sounds great. So,
first just a fun little thing, I call this Byron's law. So, the quick anecdote, one day I was working on a project
in IoT and I found myself with a hardhat climbing a ladder and
a new building construction and we're inspecting the
PLCs and HVAC equipment. It dawned on me that,
I went to school for Computer Science and I've been a software developer
most of my career. How is it that I ended up with
a hardhat crawling around? But I was loving it. So, how do you reconcile this cool new technology and
pretty critical infrastructure? So, it dawned on me. It's an interesting intersection. When you think about if you plotted coolness of a
new technology over time. So, when new technology is born, it's the latest technology, it's the latest startups. Over time, it gets
a little bit less cool. But if it hangs around
that by definition means it becomes absolutely critical
to the way we run the planet. Think about alternating
current in 1847, men want to do a fact check. I think was the Chicago World's Fair
where Westinghouse and Tesla lit up a 100,000 incandescent light bulbs with alternating
current technology. At that time, it was really cool. Everyone was like unbelievable magic
that they're seeing. Now, we just take it for granted there's alternating
current everywhere. So, over time, the cool technology
becomes mission-critical. >> Yeah. >> What we're doing with
Industrial IoT, and in fact, any of our industrial solutions, we're taking the latest technology
and applying it to the mission critical
infrastructure that runs the planet and I just
think that's unbelievable. >> Yeah. >> Right, isn't that fantastic? >> It is. It's like we keep in the cool aspect while we
work on the critical size. >> Yeah, exactly. >> It's the perfect intersection. >> Yeah. >> Yeah. So, that's
Industrial IoT for me. That's why I spend my career
doing this stuff now. Okay. So, some use cases that we see as very common with our customers, three main categories, what we'll
call smart connected products, smart connected operations,
and smart connected workers. So, a smart connected product means, I've got an asset and I want
to monitor it remotely. I want to optimize
my use of that asset. Now, I might want to detect
when it's going to fail, understand its current condition. That's very interesting and
very quick value creation. But I may also use that connectivity and make the product
more than it used to be. I might be able now to offer
the product as a service. If I can to remotely
monitor the product, understand how my
customers are using it, now I can offer that as
an on-demand service as opposed to just a piece of
gear that I sell my customers. So, there's a whole category of
use cases for Industrial IoT where the product manufacturers themselves want to make their products
smart and connected. >> It is interesting that these use cases are such in
critical pillar in the business. If you look at
the use cases that you are sharing is like we have the product, the connected product but also the connected operation and
the connected work. When we think of in this business, we are like that business will
be there to serve customers. How do they serve their customers? Through product and that product
could be a physical product or service and who is working
and innovating that product? It's the workers, right? >> That's right. >> Then, how do they do it? They do it through a set of
processes and operations. So, it's really interesting that the Industrial IoT applies
to these different pillar of business and is enabling to completely change the way
the worker is doing their job, the way the product is performing, and of course for the business, how the operations
are more efficient, we're avoiding in
the downtime and so on. >> Yeah, they really fit together. So, smart connected products. In my operations, I may have
dumb unconnected products, with our IoT technology, we can help them get those operations connected to make
them more efficient. Then of course, there are the people. We're not looking to displace people, were looking to help people
do their jobs better, safer, and more efficiently. So, that's Industrial IoT
really in a nutshell. Now, when we, Microsoft and PTC, look at the text stack
that you really need to deliver on these solutions, we see these mean capabilities, and hopefully you can
see my mouse here. So, starting left to right, obviously you've got devices that are going to be connected or connectable. There will be some
software-based edge agent that is responsible for collecting
that data through sensors or our control network and brokering
that data transmission and applying data transformation
logic perhaps some analytics on site at the edge. But ultimately,
communicating those insights or raw data to the Cloud
through a Cloud Gateway. That Cloud Gateway
is enabling storage and further processing of the data through an application
enablement platform. This is where you would build
your apps and implement your business logic and expose these insights and
experiences to end users. They'll use tools like UIs and
UXs and reporting packages. Some data will go on cold storage
for archival purposes. Some data will go into warm storage. This is operational data that you may want to read five minutes from now or tomorrow as opposed to something that you're just
archiving for posterity. Then, there will almost always be a another business system that
you want to integrate with. I might want to automate
the creation of a work order in a work order management system
like a Dynamics 365. >> Yeah. >> Yeah. >> I might want to feed an ongoing machine-learning process
where I'm constantly training AI models to make
better predictions about how my data is telling me that I have
a problem with my machinery. >> Okay. >> Yeah. Actually, that piece is very interesting because
when we talk about IoT, we think it's a new and a hyper, like customers and enterprises have been doing
Industrial IoT for years. But, what we have seen
is that previously, without the new cool technologies, all the IoT solutions were siloed and they were not
talking to each other. So, if we think for
the earlier examples, they had solutions to
manage their product, they had solutions to manage
their workers and they had solutions or systems to
manage their operation. But, these solutions were
not talking to each other. With the the newest technologies and this new reference
architecture with IoT, we are enabling
those systems whether they are old systems or new systems
to talk to each other, allowing the data to
flow seamlessly and providing the customers the
possibility to get newer insights. >> That's right. >> Yeah. >> So, that's really the, I think one of the key advantage that IoT is bringing in today
to the industry, right? >> Yeah. >> Yeah, we're connecting
the things but we're also connecting the people and
connecting the other systems. >> Exactly. To add on to that, I think for the longest time
for IoT most people were focused on the connectivity
and now there's a lot more support for
machine learning not only in the Cloud side but even
on the Edge side as well. So, we're seeing a lot
more people doing more than just connecting devices, actually doing device management
and have the true intelligence. >> You bet. >> So, Joe can you
tell us more on how PTC ThingWorx and
the different services fits in that architecture? >> You bet. So, this these gray boxes can get lit
up with some nice color. So, this previous slide
describes the capabilities. We're not talking about
any specific services from PTC or Microsoft. So, the way we laid up these
capabilities is in the following way, IoT Edge from Microsoft as the IoT Edge agent and plug-ins
from PTC that would allow that edge framework to
speak to a plethora of connected machinery through
our industrial connectivity drivers also known as our Kepware technology, to do software and content
management so that I can send calibration updates or
configuration files to my smart connected products, so it can better configure
itself and remote access, RAC remote access and tunneling, so that I can take control. When I need to do an interactive
diagnostics procedure, I have the ability to do that through a tunneled
network connection. Then moving to the Cloud, do you want to maybe talk
about Time Series Insights? >> Yeah, sure. Before that, I just wanted to add
on the remote access. I think that's a critical
thing in the industrial IoT, especially from
a security perspective. Providing remote access
from outside of the factory network is from
a security perspective like, "Hey. I don't want to do that, right. " >> Right. >> Yes. >> So, with the Azure IoT
solution like IoT Hub, we are providing and I
think works as well, we are providing the capabilities
to remotely access the devices, the machines while we respect and follow the
security best practices, such as the service assisted
communication and so on. So, those machines are
connected to the Cloud Gateway, the IoT Hub and
the IoT Hub is of course following security best practices. For example, the connectivity is only initiated by the device and so on. It provides the capability to make the data that we are
collecting from the devices, available for downstream services
to consume that data. As Joe was explaining earlier, it could be direct storage
for archival. So, that's why we have the Azure
Blob Storage but also have warm data path storage or analytics and that's where
we have Time Series Insights, which provide the capability
to also explore the data that we get from
the different devices at scale. Also, the Azure IoT Hub is connected to ThingWorx and
ThingWorx provide the capability to create different application and that's what we will
drill down a little bit further and I think we have
the demo on that as well coming up. >> Pretty cool demo coming up. So, maybe just to touch
on each one of boxes, briefly just so that you get a sense of how this tech stack
is realized again. IoT Hub helps us maintain
a secure connection to Edge devices, which are deployed in a factory environment in
an industrial facility et cetera. Blob Storage would be for
historical or archival purposes. ThingWorx is a PTC product. We call it an application
enablement platform. The role that ThingWorx plays
in this architecture is to be the one-stop shop for
understanding, my asset. So, it's real-time
operational status. We have a connection through IoT Hub, which connects to IoT Edge, which connects to the machines. But there is other
reference information that I need to understand
about my asset. I need to know
its warranty entitlement. I need to potentially know
the build of materials. What are the part numbers? If I've detected a potential
failure on a motor shaft, I'd like to know what the part
number is of the motor shaft and so I can look it up in
my parts inventory database. So, there are these other
customer business systems of record that also inform what's
going on in this solution. So, ThingWorx aggregates the 360 degree view of
everything I want to know about this asset
so that I can expose it to an end user
oriented application, in this case, Asset advisor which
we'll show a demo of as well. That implies user management. So, we connect to
Azure Active Directory which plays an important role in the
architecture in machine learning, which helps us determine
when something's going to happen if a person is not
able to configure the rule. So, if I know when a voltage
level drops below 20 volts, that's a potential problem
in a particular part of my physical architecture. That's a rule and I should be able
to configure that very easily. It's when the people don't
know what those rules are. When it's too complex or there's too much historical data that needs to be looked at to make
the determination. That's a great use for a ThingWorx Machine Learning and
Azure Machine Learning as well. So, let's take a look at
a more concrete example here. This is an example industrial asset. It's actually a conglomeration of parts that you'll typically
find in industrial machinery. So, there's a programmable
logic controller over here. This is from Allen-Bradley. Sorry, if you know what
PLCs are but if you don't, they're basically the technology
level of an Atari 2600. >> They are. So, it's
fun working with them. >> Yes. >> Yeah. >> Yeah. They're pretty basic. They're not
high-performance computing, although they do have an evolution in the near future where
they could become that. But there are really for,
"I'm going to connect some wires from buttons and electrical outlets and based on certain signals
that are coming in, I'm going to provide
certain outputs." Very basic rules. So, PLCs are often used to get
operational status of an asset. They don't do much besides just say, "Here's what's going
on with this asset." So, this PLC is wired
into an IoT Gateway, in this case a Bosch XKD. It's providing some temperature
and humidity sensor readings as well but it's also brokering our physical connection through Wi-Fi to the Cloud and it's
running an IoT Hub SDK. It's sitting on top of an electric
motor which has a drive shaft. In this case, this motor is
not doing any real work. It's not connected to any gears. It's not reciprocating anything, It's just spinning freely. But it's going to
represent a use case, where the product manufacturer of this motor rig which might be connected into any kind
of other equipment, needs to understand how they can best service and maintain
this motor setup. Okay. So, that's the setup. >> Maybe before we jump in, we have different type
of connectivity as well. This is only one example of devices
that we are able to connect to the Azure IoT Hub that we do through the SDK or also through
other protocol translation. I like your introduction
about the PLC, when we do Industrial IoT, that's all what we hear. The PLC word is like you
hear it several times. There are different types
of PLCs out in the market. Maybe if you could
touch a little bit on how we do that protocol
translation as well, now or a little bit later
that would be good. >> Yeah. We've got a little demo
architecture slide coming up. >> Perfect, yeah. >> We could talk about how
these pieces are connected together. >> Nice. Real quick, sorry. >> Sure. >> A question I get a lot is what devices are already
certified with Azure IoT? So, we do have a device catalog. So, just throwing that out but as Shafia I mentioned
and Joe too that we have the device SDKs and
Azure IoT works on all different platforms and
all different types of devices. So, you're not tied to just using
one programming language on one specific device with the operating system that
you have to buy into. >> You bet. So, let's talk
through a little story here. So, this is setup, this is our assets,
our industrial asset. It's a motor, it's got a PLC. Now, imagine that I am a service technician of
the motor manufacture, and my customers are all over
the globe and they're running versions of my motor byproduct in operational facilities
like manufacturing plants, industrial wastewater
treatment facilities may be in the HVAC system of a building like the one
that we're in right now. It's very interesting
if I can ascertain what of my customer deployed motors
have a problem today. Do I need to send
a service technician? Could I potentially make a correction over the air by sending
a configuration update et cetera. So, this is a screenshot of our asset advisor solution
for doing exactly that. In this use case, it would be for remote monitoring
and remote service. So, I can see a short list
of assets that are deployed and if I
drill down into one, I might have gotten a trouble to get notification through an email. I've clicked through
the email and now I see and as a subject matter expert
for this particular motor, I can see very clearly that there's a current fluctuation
that's not normal. There are some currents spikes. So, here's a great example
by the way of where machine learning could
inform another expert. So, I just told the story where I happen to know something about what the current signature should look
like on this particular motor, but If I'm a new technician, and by the way we do have
an aging workforce problem, where the subject matter experts for a lot of this stuff
are retiring and we have to train a new generation of service techs and
maintenance personnel. So, machine learning could have
definitely helped in this example. Maybe as next IoT deep dive
we'll connect the dots there. >> Yeah, I think that's a good one. To your point about
the workforce knowledge, that's something that I hear
consistently from our customers especially in working
with these big machines. I have customers telling
me that they have only one worker that would know how that machine works and they now. >> The machine whisperer. >> Yeah. >> Yeah. Tribal knowledge. >> Let me put my tongue on it. Yeah, its got a problem. >> I hate that multiple times. I think it's hilarious if we are from the outside but they are
living that every single day, and that's where machine learning
and the new technologies bring. How do we spread that knowledge
across the workforce instead of having only one person
as you mentioned the whisperer. >> The machine whisper. >> Exactly. You can
also see those changes in a long period of time
too, which is nice. Because you'll have parts, where
maybe it's been running for two years and it's at
its end of its life cycle. So, it's nice to be able to view those anomalies and be able to do
some predictive maintenance too. >> That right, normal
sounding when it's brand new and normal sounding when it's 10
years old might be different. >> Exactly. >> These are all the nuances
that AI can help us decode. Okay. So, in our story, back to our story. I'm the remote service tech. I'm sitting in
a call center somewhere and I'm looking at Acme Co motors. There's a problem here I need to
dispatch a service technician. Maybe I've done it
manually by clicking a button or it's been
done automatically. Now, I'm a different person. I am the service the
field service technician, I've got the hardhat on now. Wouldn't it be really
interesting if I could combine the IoT instrumentation
from this motor unit, the temperature of the censor, the current, the voltage, and combine that with some repair
step-by-step procedures. So, a fantastic combination
that we found is to use a HoloLens device or really any
other augmented reality device. HoloLens is really nice
because it's hands-free. So, here's a short video. Imagine you've got a HoloLens on, and you're the service technician. So, I'll scan
this special marker here, it's a magical little mark. I'll pause there just to describe it. Let's try that again and this time I'm not going to advance slides. So, this special mark
we call it a view mark, and what this is going to do for us, is, I'll just pause right here. It's going to tell the
augmented reality technology which is computer vision
technology essentially, where Euclidean origin
is on my AR experience. So, we know now, okay, that is motor serial number XYZ. We know how to get its IoT Hub and thing works based
data history and real time data. I know when I want to place
augmentations on the field of view, the digital artifacts that
combined with the real world view, that is augmented reality. I know where to place them, so they're perfectly placed
where the person who developed this augmented reality
experience who would have used a tool like
Vuforia Studio from PTC, where exactly they
wanted that to show up. So, that's what the mark
is doing for us. >> Again, think of the worker maybe in some cases
they've never seen that machine or that they see it
rarely or only on simulation. >> I'm new on the job, my first day. >> Exactly. >> I don't know what this thing
is, I guess it's a motor. Hand me a wrench and they told me I have to replace the motor shaft. >> Yeah. >> So, I mean, how many
times have you watch YouTube video of doing
some repair on your car? So, imagine you can have in augmented reality display that
shows you what the problem was. Let's go back up here a little bit. So, on the right of my screen
I've got my instructions, I've got a fluctuating motor current, that's what the people back
in the call center told me. I need to locate the right
part and by the way, augmented reality is highlighting
for me where the motor is, in case you didn't know
Joe, that's the motor. Then my next steps would be, what is the procedure for
taking the motor apart? I need to remove the drive shaft, I need to replace the motor itself. >> To repair this thing. >> Isn't it interesting? By the way before you even get there, isn't it interesting that
I've got the current voltage and current. Current, current. >> Yeah, sorry. >> I can verify that
the unit's turned off, it's now electrified and the
humidity level is 23 percent it's pretty dry and it's pretty
cool at 66 degrees Fahrenheit. So, the data that I would have in my remote view is also given to me in real time
and I'm hands free. >> You would show later how this data is flowing and
coming and communicating. >> We're going to peel
the covers open and you can see how this flowing through IoT Hub and into thing works and how
we realized this demo. >> One thing I would say is none
of this video has been edited. So, everything you're seeing is
actually we brought in the motor, the full hardware setup last week, we did a prerecorded. So, you're seeing a video of a surface doing AI and doing
this whole experience. So, none of this is a movie magic,
it's actually in real time. >> What you see is what you get. >> Yeah, exactly. >> This the augmented
reality experience and these animations
that you're seeing, were authored in PTC tools like PTC Creo but could be authored in any tool that does
three dimensional graphics, which is very common for product
manufacturers to produce those 3D models and animations as
they're designing their product. Now, with augment reality, there's a way to leverage
those same assets for the purposes of the on-site
training and field service. So, pretty cool stuff. >> Yeah, it is. >> Let's peel the onion. Peel all the layers of the onion
and talk about how it all works. >> We've got what you want. >> Like ogre out there. >> Yeah. >> Okay. So, once again we've spent too many time on diagram slides but, we had our remote service person who detected the problem by using
the Asset Advisor application. Asset Advisor, is based on
the ThingWorx platform which gave us the 360 degree
view of the asset. It parts information. We're able to access the animations
and CAD drawings, et cetera. ThingWorx also got it's
real-time status, the current, the voltage, the temperature,
and humidity from IoT Hub. IoT Hub, was receiving
the edge information. What was the edge information? It was the data from the PLC
and the sensor pack. It's physically on the motor, through PTCs Kepware product which is an industrial
protocol driver library. So, Allen Bradley PLC we know
that speak that language. The Bosch XDK sensor language. We know how to speak that as well. Modbus MQTT. So, basically what Kepware is doing, it's the Tower of Babel translator
for these varieties of industrial protocols something like a 150 different protocols
that we support. In normalizing that into a format that IoT Hub
can do something with. A canonicalized data format through IoT Hub processed by ThingWorx
to enable the applications. Accidentally, went to the next
video but just in time. >> Yes. >> Perfect timing. >> So, we have Kepware
sending the data to IoT Hub, and what we see here
is the VS Code which is part of the tooling that we have. >> I'm so excited to
talk about VS codes. I don't want to interrupt you. So, if haven't tried out VS code, I highly recommend it as one of
my favorite IDEs to work with now. What's great about it is, you
could have it on Windows, you can have it on Mac. I've read it on Ubuntu. It's lightweight. It's open, it's just
amazing to work with. In the last year, the tooling team here at Microsoft
spent a lot of time making it super easy to connect to all
of our services with IoT. So, one of the extensions is
the Azure IoT Hub toolkit extension, which I'm a huge fan of because
you can create IoT Hubs, you can do device management, even has IoT Edge support. So, a lot of you folks working
with IoT Hub or IoT Edge, definitely download it and try it out because it just makes the development a lot easier to work with IoT Edge. So, I just want you to do that
call out because I love VS Code. So, anytime I can do a call out. >> Well I'll give
you a plug here too. So, we built a connector
from ThingWorx to IoT Hub, and when we were building
that originally, there wasn't the
extension for VS code. So, I remember we were using
IoT Hub toolkit on the command line. That was fine. We're
cool with command lines. We're developers. But when this extension showed
up with its magical, here's your list of devices
and we were just like wow. >> It just makes it so much faster. >> This is what it's supposed to be. >> Yes, exactly. >> Yeah. >> So, I'm going to
hit play on the video. >> Yeah. Here what we see in this, we have of course, when you have your IoT Hub you connect through the
connection string and you see the different venues and
the different actions that you are able to do directly from VS code. So, you could copy
the connection string or in the use case what
we are doing is that we are monitoring the
data that is coming from that model that Joe showed earlier. So, the moral is sending
the data to IoT Hub. Through Kepware, and from the service side we are pulling that data and we are showing it here. So, in the output we are
looking at the telemetry data. The real data that
the moral is sending, and of course IoT Hub. You could send the data
in different formats, and you are able to see that
data as we scroll here over, and one comment on
the Kepware that I had also is that we are with PTC Kepware today
we did not see the IoT Edge, but there is an
integration that is in the world for Kepware on
the IoT Edge as well, right? >> Yeah. So, IoT Edge is the
perfect framework for hosting edge compute and the updating the modules that would be
performing some units of work. Kepware is fundamentally for running at the edge to do
the protocol translation. So, it's a very natural thing
for us to have a formulation of Kepware that will be in
IoT Edge container module. >> Yeah, and the cool
thing is like in VS Code you are able to see the data
and interact with the data. You are able to see the device, as well as if the device
was an edge device, we would be able to see the modules. So the Kepware modules can
interact with them and do the deployments as well
from the central way. So, what happens in
the demo that we have, is like this data is
consumed by ThingWorx, and that's what Joe will show us in how we get this data in ThingWorx and how we get it into the
screenshot shared earlier, right? >> You bet. Now of course
this podcast or video cast is an IoT deep dive. So, it's for the technical
architects and for the developers. So, we're showing you the behind the scenes view of what's happening. So, they have an understanding
of how it all fits together. It is our goal to not require
you to be parsing JSON. Obviously, end users are going
to use an application like PTCs asset adviser which
is connected to IoT Hub. But it's important for you
to have an appreciation for the hard work that's gone
into IoT Hub and ThingWorx. So, that you as
the technical architects and developers can
build on top of that. So, your customers faster to value and what they're trying
to do and not having to reinvent the will like
moving messages through an MQTT protocol bus and so
forth. That's cool stuff. >> Yeah. >> But you probably don't need
to be writing that code anymore. >> Save a lot of time. >> Yeah. >> Yeah. So, all that
is taken care of. >> It's all taken care
of so that you guys can work on the higher level stuff. The more interesting stuff. >> The coolest stuff. >> Yeah. >> The coolest stuff. So,
marching up the stack here now. We were looking at VS Code and we saw some raw messages
flowing through IoT Hub. This is ThingWorx composer
and in Composer, we're building
that view of the asset. So, we're combining data streams from IoT Hub and from
other business systems. But in this case, we're
configuring the IoT Hub itself. So, this happens to be where we
connect ThingWorx to an IoT Hub. I'm using the singular there. So, you create a hub
in the Azure portal, and you tell ThingWorx, this is where my hub is, and I can invoke as the ThingWorx platform various
services of the IoT Hub itself. So, we can configure device twins. We can configure
container modules for IoT Edge. All of the stuff that
IoT Hub exposes, it's our goal to make
that connected to ThingWorx so that you can drag things around in a configuration
interface and configure pre-built solutions
as much as possible. So, this is just a
quick view of some of the services that are exposed by IoT Hub that thing works
has an abstraction for. I'm going to zoom forward
here just a little bit, something a bit more interesting. So, now we're drilling into, I'm going to pause real quick. We've drilled into a thing. So, this is represented in
IoT Hub as a device twin. So it's reading about- I'm sure you're going to be reading
if you haven't already, background information on how
Microsoft's IoT Hub works, how PTC ThingWorx works, and IoT thing in ThingWorx
and in IoT Hub are connected. So, you can provision them
in one place and they show up in the other
and the data is shared. So, this is our connection
to the applications. Applications will be built on ThingWorx and ThingWorx
is integrated to IoT Hub. In this case, I'm
looking at my current, humidity, operational, efficiency,
temperature and voltage. So, in this small example, what I care about of this thing, in this case our motor, is its sensor data
current and humidity, temperature and voltage but also
its operational efficiencies, OEE, which is a common term
used in manufacturing. >> Some of these values, we get
them directly from the device. Some of them are calculated, right? So, some of them were
those that we were seeing on VS code earlier. >> That's right. There
was no OEE feature. >> Yeah. >> OEE is a synthesized property
and the cool thing about it is, it could have been through
a business rule that a developer or business analysts
coded in ThingWorx to say, when my manufacturing output looks like this and my asset
reliability looks like this, compute my OEE which is a magic number based on some proprietary algorithm
that my business cares about. Or it could be a prediction
for machine failures. >> Yes. >> That might have
been through, again, a simple rule or
a fancy AI algorithm. >> Yeah. >> The goal here is to normalize
that all down to the stuff I care about when I'm looking at
this asset in an application >> As in user friendly way
front operatorize, right? >> That's right. >> Exactly. >> Yeah. Now, even
speaking of persona's, the user of this interface
is not the business user. It's the system administrator. So, you don't need to be architect or a software developer to
be using this interface. You're someone who's supporting the business applications,
probably IT. And you might also be an architect, that's fine too. We're
architects here. You guys are probably architects. So, understanding how all
of this stuff is wired together is the goal
of this short video. Okay. So, the values
that we're seeing here, by the way, are the real-time values that are
coming from a connected motor. In the case of this video, I believe it was connected through a cellular network or so I heard. So, again, the connectivity, Wi-Fi or Hard line or cellular is all abstracted by the time
you get through IoT Hub. Now, that being said, as the software architects
and the system architects, you do have to be mindful about choosing the right
physical connectivity. We're seeing some five G
opportunities in manufacturing. Believe it or not, there's
plenty of wires in a factory. >> Yeah. >> But it's still
sometimes helpful to use that high bandwidth
cellular connection. >> Yeah. Then also, that's why a big reason
people are interested in IoT Edge is because you could have that extended offline support too. Just wanted to throw that in there. >> Yeah. So, that's very important. Sometimes we get
questions from people who are rather than new to
IoT and they're like, "Hey, if my things are connected, can I still see anything about it?" Like yeah, it's okay. The network
is not reliable, right? >> Yeah. >> That's the mantra. >> Exactly. >> It's cool when it's live
connected and you can interact with it right now but we're expecting
that that's not always the case. So, the most recent status and when that status was received and here's a good example where last updated. These things last
communicated on those time. We last received the
temperature value at this time. Those are important concepts because these things might be an about, and may be out of coverage. >> Oil rig or anything
like that, yeah. >> And we're providing the Azure IoT services and
on ThingWorx solutions, the capability also to get alerts or to get information
on how that device is working, like, did the device connect
from the last past five days? >> Yeah. >> We enable the capability
to use that telemetry data to see if you want to act on the device if it did not send
any data for the past five days, for example, or 30 days,
whatever, right? >> That might be the alert
just that it hasn't communicated in five days
might be the problem. >> Yeah. The other aspect, I think, is the notion of real time because we use the word real time a lot, but I think one thing that
we need to be mindful of is that the definition of real time is different from
an industry to another, from one's case to another. Sometimes it's one second
some people would say is real time. Sometimes it needs to be nanoseconds. >> Great. >> Yeah. >> A real-time operating system
doesn't multitask. >> Yeah. >> Exactly. >> It probably doesn't have an IP stack because
it's mission critical. >> Yes. >> This would be a PLC controlling a safety solenoid on a motor that's maybe bringing
a blade down, right? That's an example of a
real-time operating system. When the button's pressed, stop the blade. That's real time. >> That's real time, yes. >> Five milliseconds,
100 milliseconds is not real time. >> Yeah. >> Yeah. >> Yeah, that's a great point. That's where the cool in
the critical interests. >> Yes. >> They have a right there.
There's an example of that. Okay, so let's see
where are we are now. So, we actually already did this one. Okay. So, another
important aspect of IoT, we've been talking a lot
about data receiving. So, ingestion of data from machines sensors and so
forth so that we can trigger alerts and look at
the data flow et cetera. We also not as commonly but importantly want to sometimes
send commands to things. We might receive a million values
a day from a connected motor and once in a blue moon literally we need to send it a software update. But when it's time to
send a software update, we need to be able to do it reliably. So, IoT Hub has a bidirectional
communication to the asset. In this case, the edge agent that was running on our Bosch XKD
and we're able to control through Kepware
and its connection to the PLC physical
actuation of the motor. It's going to be a little
hard to see in the thumbnail. So, bear with me. I'll
skip forward over here. Again, this is the system
administrator view not a end-user application. But we're going to
execute a remote command. Click the "Execute" button. That you can see. It's
really hard to see. >> Very tiny. >> Trust us, if you're in front of the motor, you would have seen it. The solenoids, there are
two solenoids attached to. Maybe I'll just quickly go back to and I'll talk through this instead of trying
to look at the tiny video. >> Yeah, a big reason why we prerecorded was because
this motor is pretty loud. So, we wanted to make
sure that you guys weren't just hearing this loud motor
running the entire time. >> Which by the way is pretty
common in the industry. >> Exactly. >> Places are allowed. >> We have to make a decision between you guys hearing us, so the model. >> Yes. >> Yeah. >> Yes. So, again this motor, we want to slides and load. This motor had if you recall a PLC, this motor has a PLC which is connected electrically
to the solenoids. The PLC is, as we talked
about, a real-time OS, it doesn't have an IP stack, it doesn't know how to do
cloud-based communications. So, that's happening
on this Wi-Fi module. We're sending a command to the Wi-Fi module and telling the software agent
running on that module, talk to the PLC and asked the PLC to set its output for solenoid
it energized to on. So, that's how it works
behind the scenes. Kepware knows how to talk to the PLC, the Edge module knows
how to talk to IoT Hub. So, it's brokering that connection between the physical asset and it's actuation through the PLC and
electronics and electrical servos, in this case a solenoid and
giving it an energized command. So, that's Industrial IoT for you. Certainly not the beginning or
the end but maybe a slice of the world that we live
in and the world that our customers have been
implementing solutions in. >> I want to do a shout out
to all of our folks on chat. Thank you for helping
out to answer questions. They've done such a great job that I haven't had to ask any questions, as the host publicly. So, thank you, Carl, Mustafa, and Olivier
and others on chat. It's super helpful for
you guys to be there. So, since we don't have any questions because
they were all answered, I think we can go ahead and jump into that last slide to
talk about resources. >> Okay. Actually. >> Go ahead. >> Because it feel a
little a [inaudible]. Maybe Sophia and I can talk about
what we see on the horizon? >> Perfect. Yeah. >> So, we've done a lot
of work together and combining our technologies in
support of solutions like these. So, I'll interview you
for a moment, Sophia. What do you see as the next critical cool capability that intersects with critical needs, in the use cases that
you've been seeing? >> I think in the industry, the one cool thing that I would expect to boom in
the next few years is the intersection between
the IoT and the simulation. We tend to talk about it like we
hear all about the digital twins, which is the digital representation of a device or a physical piece. But in IoT, one thing, one, we are a company that building new products is sometimes we
don't have all those data. Sometimes we don't get
all the context about the product, and what I'm building
a very expensive new product, I want to simulate
that missing piece of information. So, the thing is like
how do I do that, how do I bring the simulation
aspect to the physical, the context data that I get and I build a real digital twin
for that device. So, I think there will be a very interesting
innovation around that. I'm hearing a lot of requests and interests from
customers in that space. >> Yeah. That's really, really interesting and
we have a partner Ansys, who is working on this and we
are collaborating with them. So, just for everyone
to get up to speed, there have been for quite a number
of decades, in fact, the activity of building computational fluid
dynamics-based simulations of physical products during
the design of the product. So, I want to understand the finite element analysis
and mechanical phenomenon, the physics-based phenomenon, that my design will undergo with the forces and
temperature and et cetera, thermal characteristics
over the shells, et cetera. So, there are companies that have specialized in
such mathematical models, Ansys is an example. So, finite element analysis
computational fluid dynamics. The thing is it takes a while
to run those calculations. So, it's good for
the design activity. Here's my design. Maybe here's the CAD model, I'll dial-in with
PTC Creo and I'll build a computational model of the thermal characteristics
and force characteristics, run that simulation, go have lunch. It's not that slow,
but it's still not real-time or hasn't been real time. So, here's a cool stuff
that's happened, machine-learning happens to
be very matrix oriented, which it's processing
numerical matrices. The collective technology world, had the insight, not too long ago, that a GPU is fantastic at matrix multiplication for calculating
three-dimensional graphics. So, using GPUs for AI, fantastic. It also turns out that the simulation stuff is
also matrix mathematics. So, companies like Ansys
have been adapting their simulation models and calculations to take
advantage of GPUs. It's gotten so fast and
near real-time that we can imagine having that simulation, which was already built when
the product was designed, deployed at run-time and actively looking for anomalies
based on the simulation. Sometimes machine learning
and AI is a good fit, sometimes augmenting with
the machine learning is doing with those pre-built simulations fills in a missing piece of data that
the sensors may not capture. >> That's exactly. >> The coolest thing is that we
bring that to the Edge's rule. That's the work that is
ongoing. Like the market. >> Yeah. Wait where it is to happen. >> Yeah, exactly. So not we don't need anymore all that compute power to be in
the cloud or centralized ways, we are getting that closer, as you said, where
it needs to happen. So, are we good to close or
still have something else? >> Yeah. I think we're good to close. I think you're totally great on
that digital twins and [inaudible] , just where we're going to see a lot more information and
more folks doing more support there. So, for resources, we have
a couple of links for you. One is the PTC
Microsoft landing page. I want to make sure to upload
these slides on our tech community. If you haven't gone and checked
out our tech community's site, we have one specially for IoT, so I'll make sure to
upload the slides there. PTC also has a developer
portal homepage. It's great for tutorials, how to get started, all of that. They're also developing. If you guys haven't
played with the MXChip, they're developing a lab
on how to work with ThingWorx and Azure IoT together. So, this MXChip DevKit, you can get it at Seeed Studio, you can get it in a lot
of different places. But it's nice because if you don't have a hardware engineering
background, it's perfectly fine. I do and I still really
love this device. You get connected right away, connected through IoT Hub,
working with ThingWorx. So, the lab they're still developing, but in a couple of weeks, if you go to that
developer portal homepage, they'll have that lab for
those folks who have this kit or wanting to get this kit
to quickly get started. >> Yeah. By the way, I
went through the lab myself last week with an MXChip. Even for people like us, [inaudible] it's pretty cool, it helps you especially
if you're new to IoT, it helps you understand the concepts. What is the role for the Edge agent? How does IoT Hub really work? The stuff that we've clicked through the screens and videos
we showed today, you can get hands-on
with that stuff and get an appreciation for it and understand where as you're implementing projects for customers or your own companies, where your value should be and how you should
interact with these tools. >> Exactly. What's nice it also has a lot of sensors
already on there, so you don't have to worry
about breadboarding, a bunch of sensors to
the device just so you can. >> No juggle wires. >> No wire wrapping, none of that to work with the device. You can really get started
on the cloud side. Then we have an upcoming event,
so Hannover Messe. If you haven't been to that event, it's one of the largest events
for Industry 4.0, everyone's there who's part
of that industrial industry. So, I highly recommend
checking out PTC there, Microsoft would be there as well. Chat with us and we would love
to meet up with you, guys. >> I will be myself there, so feel free to reach out, we'd be happy to meet in-person. >> All right. Great. >> All right. Well,
thank you so much for joining on our first IoT Deep Dive. We're going to be doing these on
Wednesday, every other Wednesday. So, definitely, if you have
ideas for future deep dives, anything that you really
want training on, let me know. Thank you, guys. >> Thank you. >> Thank you. >> Thanks. >> Bye. >> Bye-bye. >> [MUSIC]
