[MUSIC] >> Hi everyone. My
name is Pamela Cortez. I am a Senior PM and an advocate here on the
Azure IoT Product team. I'm excited to be here with Cory, who's awesome and amazing to go
ahead and speak about Azure IoT. >> Hi Pamela. Hey everyone. I'm Cory Newton-Smith, and I'm a Principal Program Manager on
the Azure IoT Central Team. >> For this Ignite session, we're going to be
talking about building end-to-end IoT solutions secured from the Edge to the
Cloud with Azure IoT. First, we're going to go through the two paths for
building IoT solutions. Next, we're going to go ahead
and go through that first path, which is all about
building from the ground up with our Azure services for IoT. Then for the second
half of the session, we're going to be talking
about building with a fully managed IoT
application platform, which is Azure IoT Central. Then we're going to
end with giving you some resources on how you can
get started and learn more. Let's go ahead and jump into it. Let's go over that overview of those main options for building
an end-to-end IoT solution. The Azure IoT portfolio
consists of a number of well-composed entry points for
building these IoT solutions. At the heart of it
and the foundation of the stack includes our offerings for device support coupled with an extensive set of Azure
platform services for IoT. To work with those services, we created a guide
called the Azure IoT Reference Architecture that documents how to use these building
blocks and how they come together to
support IoT solutions. We're going to go over
that in just a moment. Moving up the stack, we have our Azure Digital Twins. You can use that to model and interact with any
physical environment, track the past, and
predict the future. Then at the top, our
highest level entry point for solution building is with our fully managed
IoT application platform, Azure IoT Central. Crossing these different entry
points for building IoT solutions, we have the Azure
Security Center for IoT, which is an Azure-based
Cloud security service providing a comprehensive IoT
device threat protection. You can secure your
entire IoT solution from device to Azure Cloud, and then after that, you're
full end-to-end solution with the Azure IoT portfolio
and other Azure services. >> Let's talk a little bit
about how our investments in IoT have evolved over
the last several years. Five, six years ago, most of our R&D investments were really targeted at the early market, people who wanted the newest things. But as IoT as a technology has evolved across or through
the adoption life cycle, we really find
ourselves in this chasm between early market
and mainstream market. In order to address
the mainstream market, we really have to recognize
that people will be looking for more complete
solutions and convenience. Let's bring these two
dialogues together. Our investments in
Azure platform services has really empowered
the early market. In fact, in just a second, Pamela is going to demonstrate how to build a solution from the ground
up using platform services. In order to address
the mainstream market, we really had to create a number
of higher-level entry points. You see that in the Azure
Digital Twins offering, and also in our IoT Central, which I'll demonstrate later. But before jumping out
to the demonstrations, I want to spend a few minutes
talking about how you might find the right
solution for your business. We found that this decision is often influenced by a number
of important dimensions. The first of those
dimensions is expertise. How much proficiency do you have
in building Cloud solutions? The second is control. What aspects of the solution do
you want and/or need to customize? The third one is price. How important is it
for you to be able to fine-tune the price of your solution versus having
price predictability? The strength of your answers to these questions will really sway
the decision that you make. Answers to the left of the
middle will typically indicate that you want or need a more
platform services-based approach. Sure, this will require the
greatest degree of expertise, but it will also give you the
most control and configurability. If you find that your answers
put you to the right of middle, you'll probably end up with a more
managed services-based approach like that of Azure IoT Central. You won't have to spend an extensive amount of time
doing system architecture, and sure, you'll lose a
little bit of that control. But you'll get, in exchange, price predictability as well
as [inaudible] to market. From here, I'm going to
hand it over to Pamela, who's going to
demonstrate how to build an end-to-end solution using Azure. >> Thanks Cory for that overview. Now let's walk through how
to build from the ground up with our Azure services for IoT. High level, building with Azure platform services is like
building with Lego blocks. You can treat the Azure services for IoT as individual blocks
that can work together to help you build your custom
end-to-end IoT solution based on your business
needs and scenario. Now, different companies have
different business needs. IoT solutions can span from many different industries
and use cases. How you build an IoT solution
and which services you should use can look different from other
IoT solution architectures. With that said, most
IoT solutions have certain key components
that are important and common across a majority
of IoT solutions. To help you with your journey of
understanding those components, we've created the Azure IoT
Reference Architecture guide. Created for both beginners
and seasoned pros, this resource will help you learn about what are those
different core components and which Azure services for IoT can help you start building
your IoT solution. Before we get to the demo, let's review high level, the common architecture
for IoT solutions. The core components of
this architecture can be broken down into things,
insights, and actions. Our things are IoT devices
which can communicate either directly to an Azure IoT Hub with
the help of our device SDKs, or by means of Edge devices. Azure IoT Hub acts like the Cloud gateway for these
devices connecting to Azure. Once data arrives to Azure, we can begin to process, identify patterns, and
operate on those insights. We can use services like
Azure Stream Analytics to filter relevant information
where the data is [inaudible]. This allows us to instruct time sensitive insights into warm
path storage for immediate use, or offload data into cold
storage for long-term processes. Once data's in the Cloud, scalable integration of that data to a line of business applications enables the ability to take action on the data
produced by our devices. Now, let's take a look at a solution architecture
for a popular scenario, workplace health and safety. Then we'll walk through
the demo of this solution. Workplace safety is a popular
scenario because companies want to know the tools and training that they're putting
in place are working, and that they can respond quickly and appropriately to safety issues. You can see that this
end-to-end IoT solution follows that common workflow of
things, insights, and actions. Now there's two parts
of this solution. The first part is site safety. This will be implemented using
Azure Maps and geofencing to alert employees or equipment are in a physical location that creates
an unsafe work environment. We'll track the workers location
by the use of their cell phone, allowing us to determine where in
relation to a defined geofence. This allows us to alert
the supervisor when the worker is found to be in a
location that they shouldn't be in. The second part of this
scenario is checking personal protection equipment so PPE. Workers who enter the work
site are scanned by a camera, which is used to ensure
proper safety gear is worn. If not, we can produce
a notification to alert the current shift supervisor of
the potential safety violation. Now, let's see this first
part of the demo in action, which is the site safety
part of the solution. We'll be using an Outlook
account to receive notifications when a
worker has entered a location that they're
not authorized to be in or a zone that we have set
as potentially hazardous. You can see that we have an Outlook
account already ready to go, that we will be checking out if
the notifications were sent. First, we're going to set up a geofence by using Azure
Maps geofencing service. Azure Maps is a collection of geospatial services and SDKs that use fresh mapping data to provide geographic context to web
and mobile applications. Now that the geofence has been set, we will showcase a worker going
into the unauthorized location. You can see that I'm going to go to my Azure Maps demo application. Normally, this is run on the
web or on a cell phone app. Now, the worker has gone
into this location, we should start seeing an alert. So you can see the alert
has been triggered, and it shows that I have entered the construction zone
without authorization. But the real power here is
actually seeing this in Outlook. We are using the Azure
Maps Integration with Azure Event Grid to send event notifications to other services like Outlook and trigger
downstream processes. This allows us to react to
crucial events in a reliable, scalable, and secure manner. We can see the alert
in the e-mail now. Megan, our Site
Supervisor can now see someone has entered an
authorized location and now, she can take action quickly. Now, let's go through the
second part of the solution, which is checking to
see if our workers are wearing their personal
protection equipment. In this case, we're
going to be checking if our staff are wearing
an appropriate mask. We'll do this by
deploying an AI model to our camera that is running
the Azure IoT Edge runtime. First, we need to create the
model for object detection, and we're going to be
doing this by using Cognitive Services, Custom Vision. Right now, we're in the
Custom Vision portal. We've already started training our
Computer Vision model by simply uploading images to
help us identify what the workers look like when
they're wearing mask and without. Now, let's label a few images. The model tests itself on these and continuously improves
precision through a feedback loop when
we add new images. To tagger images, we simply click on the image that I uploaded
earlier and add tags. As you can see, I clicked on Contact, clicked on an image which
I am wearing my mask, so I'm going to tag it as
Person wearing their PPE. Then, I'm going to go
ahead and hit "Save". I might do that to the last image. You can see I'm not wearing a mask. I'm going to tag it differently, A person with no PPE. Then, I can go ahead and hit "Save". You can see that all images have been tagged and we are ready to go. Now, all the images are
tagged with just one click, we can export our train safety models to our devices and in this case, our devices and camera. This way, our device can
do all the processing locally and only send out the
information that is needed. Now, there's a lot of
different ways that you can deploy the model
actually on the device. For this solution, I am using the Azure IoT Tooling
extension support for VS Code. Now that we have the model, we're going to need to deploy
it on our Edge device. For this solution, I am leveraging our Azure IoT extension
support and VS Code. This makes it easy to
setup our Edge device, deploy our safety AI model, and get everything up and running. If you install the Azure
IoT Tools extension pack, you get all the extensions to help you get started with Azure IoT Hub, Azure IoT Edge, and other
services for Azure IoT. Now that the model is
deployed on the device, I'll pull up an earlier live
video-stream from our camera. In the camera stream, you can see that I'm holding
up another AI camera to showcase what the Azure IoT
Starter Kit looks like. This is a great kit that we partnered with Qualcomm
to showcase and teach anyone how to get
started with IoT on the Edge. You can see when I put on my mask, it now says, "A person is wearing their personal
protection equipment". When I take off my mask, it says, "I am no longer
wearing my mask". When I took my mask off, it should have triggered an alert. So if we go back to
our Outlook account, you can see that an e-mail
has been triggered. So it did deliver that e-mail
notification that a person has been identified with no personal
protective equipment and what the camera was, which this was my camera. It was the AI DevKit and so
it even gives a timestamp, very similar to the first
part of the solution. Let's use Time Series Insights
to visualize the data coming in. What's great about Time Series
Insights is a nice way to be able to gather insights so you
can later on take action. We have our solution setup and let's go ahead and find our camera and
the workplace safety solution. I'm going to go ahead
select Camera number 3. This is our camera that we
have been using for this demo. I want to visualize when a person
is not wearing their mask. What's nice about Time
Series Insights is that you can visualize
data from the last hour, or days, or months, or years, you get to
set the time span. For this particular solution, you can see that green is
when I was wearing my mask, and then, red is when I wasn't wearing my mask and when
an alert was triggered. This can be really helpful
to see this data this way because for this
particular solution, we can see that all of these triggers happen at
a certain amount of time. So you can imagine if you're the Site Supervisor
and you notice that one shift tends to get more
incidence than others, maybe more training is required
to help lower the number of potential safety violations That was the full demo for the workplace
health and safety architecture. If you're interested in doing
this solution yourself, stay tuned to the very
end where I'll talk about where to get a
hold of this tutorial. Now, I'm going to go ahead
and hand it off to Chloe to talk about how to build
with Azure IoT Central. >> Thanks Pamela. But
before jumping out to do demonstration of
Azure IoT Central, I want to review again the different options for
building IoT solutions. You just saw Pamela gave
you a demonstration of how you build a solution
using Paths Services, which is Option 2 on the slide. Now, I'm about to jump
out and show you some of the latest features
in Azure IoT Central, which is a fully managed
application platform. It's really built to simplify the connectivity and monitoring
of IoT devices so that the IoT data coming out
of those devices can be integrated into your
downstream business systems. This is IoT Central. I'm going to start off by clicking "Build" here in the left navigation. In IoT Central, you can get
started in a matter of seconds. You can either use a
custom application or one of our industry
application templates. In fact, just this month we released
the video analytics template, an enrich end-to-end
tutorial to help you set up an Azure IoT Edge live
video analytics solution, leveraging IoT Central
and Azure Media Services. Please check out our docs and latest IoT Show for a complete walk through. Today, I'm going to
walk you through some of the newest platform capabilities using MyIgnite Asset Monitoring
Demo Application right here. Let's start out with device
connectivity and look at how IoT Central simplifies connectivity by starting with our
device templates. There are several device templates
defined in my application. A device template is a blueprint that defines the characteristics
and behaviors of a type of device that connects to the Azure IoT Central application. As a device agnostic platform, IoT Central can really manage devices across a
really broad spectrum. Everything from micro-controllers
like Azure Sphere, to devices running on Azure IoT SDKs, to Edge devices that
aggregate, process, and provide gateway capabilities
for other IoT endpoints. As a managed platform, IoT Central ships
features every month, and we're super keen to hear
an act on your feedback. In fact, just a few months ago, we released a tutorial
detailing how to connect in Azure Sphere
DevKit to IoT Central. If I Click "New," you can see support for the Azure
Sphere device template. I used this earlier to connect
a sphere device for the demo. Let's go have a look over in Devices. This Sphere device is reporting
temperature values in a real-time dashboard that are defined in the view and the
Sphere device template. To assist with modeling and
on-boarding your devices of all types, we now support raw data view. This provides visibility into
both modeled and unmodeled data being sent to your device
and received by IoT Central. Asset monitoring is a
really common IoT scenario. Today, I'm going to
use this solution to show you how IoT Central can simplify the connectivity in management of multiple devices in a single
application instance. Let's start with an
overview of my scenario. I'm a solution builder providing an asset monitoring
solution to two of my customers, Fabrikam and Contoso. Both perishable goods providers they are interested in
two monitoring scenarios. The first scenario is related to refrigeration units
within their warehouses. The second is the tracking
of connected shipment bins that they use for transporting
their goods to distributors. Now, let me show you how the
IoT Edge Gateways here in this picture are on-boarded
into IoT Central. Here's the device
templates we have modeled, both the IoT Edge
asset-tracking gateway, as well as its downstream environment
sensors. Lets start there. We've imported the Rigado
S1 Sensor Template from the device catalog. This is a simple, waterproof, portable temperature
and humidity sensor. You can see this in
the model telemetry. Here we have the IoT
Edge gateway template. It has relationships
created for the S1 sensors. These will become downstream
sensors for these Edge Gateways. It also has three IoT Edge
runtime modules defined. The first one is for
aggregating temp and humidity across the
downstream devices. The second one is actually
a Stream Analytics modules that uses windowing
technique for gathering temperature and resetting
the devices if they cross the 70 degree mark
for over five minutes. The last one is an Azure function
module deployed on the Edge. This template also has
a few muse defined for visualization of data
and customer details. These published templates are used
by devices in the application. Let's go take a look at the
devices in the Devices board. This is a gateway deployed for my customer Contoso and building 17. These are the custom views defined in the device template
we looked at earlier. Default aggregate tiles
and time series graphs. Let me see the gateways
device metrics. Over here in modules, we see the IoT Edge
modules being hosted by the IoT Edge runtime
and their current status. In the Manage tab, you
can see Properties. Here we can increase the frequency of data transmission in
the aggregator module. In downstream devices, I can
see the sensor relationships. Looks like this gateway is coalescing information from 10
different cooler sensors. Let's jump out to our application
dashboard and see how we can monitor both the gateway as well as the cooler sensors
connected downstream. Here you can see the Edge
gateway aggregations for temperature and humidity. Here you can see the downstream
coolers sensor data. Now that you've seen
different types of devices connected to IoT Central, I'd like to show you how our jobs capabilities can help
you manage these devices. Jobs can be used to configure command and even update
devices at scale. Here you can see a job I
ran previously that sends a command to just under
1,000 logistics gateways. This job ran in just under 2.5 minutes and have one failing device. Let me show you how I
can update a property on devices at scale in
just a few clicks. Let's go ahead and perform this
update on the logistics gateway. Remember that these gateways are attached to the
shipment bins and have a number of sensors like tilt to ensure that contents
aren't damaged, and light sensors to ensure that containers aren't
unnecessarily breached, which could indicate
damage or even theft. Let's update the light sensor
threshold on these devices to ensure that they're uniformly
set with a low tolerance, because we don't expect our
containers to be open during transit. You can see that this job is
completed and it's more than 1,700 bins and the system spanning across both my customers in
about half a minute. To wrap things up,
I'd like to show you just how easy it is to
take the device data in IoT Central and integrate it with other workflows using our new
data export capabilities. Here you can see
data-sharing happening with both my customers as well as
my logistic subcontractors. These investments and export
that you filter enrich data as it's sent to
destinations like Event Hubs, Service Bus, Webhooks, and storage. I set up a few destinations
already in my subscription. Let's set up an export
for my customer Contoso. Both of my customers actually wanted to enable quarterly business reports joining their IoT data with data
from their CRM and ERP systems. I really want to make
sure that I only share the data that is related
to their logistics roots. Using a Customer Name filter, I easily set up an export to send all device telemetry from Contoso shipment bins to
their storage account. Let me quickly do the
same for Fabrikam. If you remember earlier, I mentioned that he used several
subcontractors for shipping. They use an application
to scan the bins when they're picking them up
from Contoso or Fabrikam. This application uses
the IoT Central API to write a crowd property for the
shipment bin with their name. Northwind Traders would like to receive a stream of
device geometry for these bins to their Event Hub
endpoint so that they can integrate IoT sensor data into
their business application. Let's go configure this export. Northwind has already shared their event hub endpoint
connection information with me, and I've gone ahead and
set up a destination. Now let's create an
export that filters telemetry data from devices
where Northwind is the shipper. Let's also enrich these messages
with the customer name property, since Northwind is a
logistic subcontractor for both Contoso and Fabrikam. There you have it. Three new
data exports are configured in three different down stream
consumers with pre-filter and enrich streams for use in
their own business integrations. These are just a few of
our ongoing investments in our managed application
platform IoT Central. Our mission is to be the easiest and most
cost-effective way to connect and manage
IoT devices at scale, enabling the use of IoT data
at scale in your business. You can always find the
latest updates for all of Azure services on the Azure
Update sites in here. I've included a few
of the latest updates to the Azure IoT Central platform. As you can see, we're continuing to improve and iterate based
off of customers feedback. We'll leave you with this. At Microsoft, it's our
mission to empower every person and every business
on the planet to achieve more. In IoT we've really been working
to simplify IoT connectivity and monitoring so they can continue to see value realized
along this spectrum. Thank you so much Pamela, today, for allowing me to
come and talk with you about the different options for
building IoT solution on Azure. >> Thank you so much
Cory for going over how to build with Azure IoT Central. It was great to speak with
you today here at Ignite. A huge thank you for
everyone who's watching. I hope everyone is staying safe and enjoys the rest
of Ignite as well. Before we head out, we want to share with you some
resources on how to get started. The first resource
is Microsoft Learn. Microsoft Learn is a great way to upscale across our Azure portfolio. We have a collection of tutorials and hands-on labs for both Azure IoT central and our Azure IoT platform
services, including IoT Hub. What I love about Microsoft
Learn it's really for everyone. If you're a developer, or IT Pro, or a business user, there's going to be tutorials and training available
to help you upscale. Our next resource is a Ignite
IoT announcement blog post. This is where we're
going to list all of the announcements and links
on where to go to learn more. If you're a developer watching now, we really recommend you checking out our IoT developer
resources blog post. We're going to list all of
these great resources for developers developing IoT solutions. There's includes upcoming events, the Azure IoT architectural
reference guide, the tutorial on how to create your own workplace
in health and safety solution that we
demoed here today and many more developer resources. Our last resource is the Deep Dives. Deep dives are interactive training
live events for developers, architects, or anyone
building IoT solutions. Microsoft engineers
and guest speakers like Cory, myself and others, do these technical deep dives about a new feature or scenario
at these events. Thank you again for
everyone who's tuning in and we really hope that
you have a great day.
