Table of Contents

 

Back to top


Introduction

When you create a "Remote Monitoring" IoT Suite solution, what you actually get is a whole bunch of cool technology wrapped up in one bundle. Everything you need to configure and end-to-end solution for device data ingress, monitoring and control.

IoT Suite is an initial set-up wizard, to help your business take all the pain out of spinning up a proof-of-concept, or the initial architecture for your project. It will configure all the parts and wiring that you will need to create the whole solution.

This is great for spinning up "proof of concept" example, and even to lay the ground work for a final production solution You could just replace the dummy devices with real ones, and define your own rules. It is also great for generating dummy data, if you are working on just a user interface tool.

Because the solutions are available on GitHub, you can trace all the parts and wiring underlying the whole end-to-end solution and scenario. 

This is just one example of how you can set up the system. Once you understand everything, you can chop and change as suits your needs with each project.

This article is a detailed checklist and breakdown of each component you get "out of the box" when you use the IoT Suite tool to create your IoT solution. 

For a detailed description of adding devices to this solution, read IoT Suite - Remote Monitoring - Adding Live and Simulated Devices.

 

Back to top


Overview

 

Below is the [to date] published diagram from Microsoft, which shows the basic configuration, as shown here:

 

Below is the actual list of what you get by default...

 

Any resources created in "IoT Suite" will be grouped into a Resource Group named after the solution name you chose at the start of the process. All items of the group can therefore be easily found and maintained using the Azure Portal blades.

Note: The main list does not show everything. You have to click the box header or the "..." to see the full list blade.

Back to top


IoT Suite Components 

I have matched and documented all these resources in this article.

The name of each feature of the solution above is highlighted below in bold and red.

Names of the actual parts or each resource are also highlighted in bold.

Azure IoT Hub

Name: ThisIsDemo

Hostname – ThisIsDemo.azure-devices.net

<Image to come>

This is the "public hub" (as opposed to the "internal" Event Hub detailed below) that devices connect to for data and return commands.

Settings

Shared Access Policies

Five policies are created, below are the main three:

  1. iothubowner – full permission
  2. service – service connect permission
  3. device – Device connect permission

Messaging

Pricing & scale

Users

  1. Subscription admins – The Azure subscription admins and co-admins

DocumentDb

Name: ThisIsDemo-iotsuite

This holds the master-list of registered devices and their differing properties. You can edit these directly in Azure Portal. Each device is a separate record in the database.

Storing "device provisioning" data in DocumentDb means all devices are in one place, but can have their own "schema" of different fields. So each document is device type dependent.

 

<Image to come>

 

Below is an example of a device file (document) with a few emboldened points of interest:

{
"DeviceProperties": {
"DeviceID": "SampleDevice003_278",
"HubEnabledState": true,
"CreatedTime": "2015-10-03T14:01:10.0018032Z",
"DeviceState": "normal",
"UpdatedTime": "2015-11-05T02:33:19.9573117Z",
"Manufacturer": "Contoso Inc.",
"ModelNumber": "MD-81",
"SerialNumber": "SER5553",
"FirmwareVersion": "1.30",
"Platform": "Plat-81",
"Processor": "i3-4792",
"InstalledRAM": "43 MB",
"Latitude": 47.659159,
"Longitude": -122.141515
},
"Commands": [
{
"Name": "PingDevice",
"Parameters": null
},
{
"Name": "StartTelemetry",
"Parameters": null
},
{
"Name": "StopTelemetry",
"Parameters": null
 
},
{
"Name": "ChangeSetPointTemp",
 
"Parameters": [
{
"Name": "SetPointTemp",
 
"Type": "double"
}
]},
{
"Name": "DiagnosticTelemetry",
"Parameters": [
 
{
"Name": "Active",
"Type": "boolean"
 
}
]},
{
"Name": "ChangeDeviceState",
"Parameters": [
{
 
"Name": "DeviceState",
"Type": "string"
}
]}
],
"CommandHistory": [],
"IsSimulatedDevice": 1,
"Version": "1.0",
"ObjectType": "DeviceInfo",
"id": "3ad1a488-17d7-48cf-bba1-de65dcd5dc08"
}

 

The "id" on the last line is the actual document name, IoT Suite creates this as a GUID so that every file is unique, even if device names are the same.

Azure Event Hubs

 

Name: ThisIsDemo-servicebus

Type = Messaging

Two Event Hubs are created beneath the IoT Hub, for splitting the device data into separate streams for active monitoring applications and acting on events. These "sub-hubs" take filtered sub-sets of data, without commands or rules.

This data streams are therefore cleaned up of unrelated data and the multi-traffic complexities of the two-way IoT Hub. This filted data tree structure also means IoT Hub itself is not overloaded with too many clients (listeners).

This solution also means it [critically] remains fast and responsive to the deluge of data. Nothing is lost, ignored, or unresponsive. Queues don't overflow.

Responsiveness is critical on the device/gateway side too, so it isn't locked up on a slow event hub call and missing a change or trigger on it's side.

 

Events Hub

Name = ThisIsDemo-ehout

This is the hub which monitoring applications connect to, for a live data feed. It is a sub-set of the messages from IoT Hub. Only messages that relate specifically to device data. It is populated from the ThisIsDemo-DeviceInfo Streaming Analytics job. 

 

Rule Alerts Hub

Name = ThisIsDemo-ehruleout

This hub is where device data is copied to, from Streaming Analytics job ThisIsDemo-Rules if it breaks any of the rules, which are defined in the devicerules blob. This provides a tidied and isolated data set, for a scalable number of clients (like web users) to connect to.

 

Storage

Name: thisisdemo

This is where history data and system configuration files are kept. This shows just one way you can configure your IoT solution. You could for example use SQL for provisioning, or stream your data into Table Storage, which is quick and cheap and has just enough indexing (Partition Key and "compound" RowKey) to analyze for reporting tools.

In this default IoT Suite configuration, Microsoft use Blob Storage, with date-split directories and files defined in the Streaming Analytics "path/pattern" field.

Table Storage

 

Name = DeviceList

This table holds a list of which devices are to be simulated devices.

It holds the basic information the back end simulator needs to act on behalf of the device.

Columns:

Name = DeviceRulesNormalizedTable

Columns:

 

Blob Storage

 

Container: actionmappings

Below is the content:

[{"RuleOutput":"AlarmTemp","ActionId":"Send Message"},{"RuleOutput":"AlarmHumidity","ActionId":"Raise Alarm"}]

Container: devicerules

Below is an example of the content:

 

[{"DeviceId":"SampleDevice001_278","Temperature":38.0,"Humidity":48.0,"TemperatureRuleOutput":"AlarmTemp","HumidityRuleOutput":"AlarmHumidity"},{"DeviceId":"SampleDevice001_510","Temperature":38.0,"Humidity":48.0,"TemperatureRuleOutput":"AlarmTemp","HumidityRuleOutput":"AlarmHumidity"}]

 Note, the value shown above is the limit that breaks the rule.

 

Container: devicetelemetry 

deviceid,averagehumidity,minimumhumidity,maxhumidity,timeframeminutes
SampleDevice002_510,30.806544133199651,20.0014513949869,49.867387465139473,5
SampleDevice002_510,30.662442867577607,20.0014513949869,49.867387465139473,5
DeviceId,Temperature,Humidity,ExternalTemperature,EventProcessedUtcTime,PartitionId,EventEnqueuedUtcTime
SampleDevice002_510,33.544336569190349,25.443365691901725,,2015-10-03T18:00:15.4907064Z,3,2015-10-03T16:58:13.8540000Z
SampleDevice001_278,33.955065882557506,29.550658825575642,,2015-10-03T18:00:05.7871002Z,0,2015-10-03T16:58:14.4620000Z

 

Container: thisisdemo-ehout

 

Container: rulesoutput

deviceid,tempreading,tempthreshold,humidityreading,humiditythreshold,temperatureruleoutput,humidityruleoutput,time
SampleDevice001_510,35.849473366071265,38,48.494733660712456,48,AlarmTemp,AlarmHumidity,2015-11-01T00:00:07.7290000Z
SampleDevice001_278,35.849473366071265,38,48.494733660712456,48,AlarmTemp,AlarmHumidity,2015-11-01T00:00:13.5030000Z

 

Stream Analytics

IoT Suite creates three SA jobs that take the data from IoT Hub into three streams of data:

  1. Device Telemetry – Raw and normalized data stored straight into blob storage
  2. Device Info – For monitoring clients like the Web App
  3. Device Rules – Any data that breaks the rules you define in the rules blob file

 

#1 Name: ThisIsDemo-Telemetry

The input source is your IoT Hub.

This is a two way hub, so first only device ingress data is filtered, using a "WITH" statement to create a subset of traffic in a temporary table.

It then does two things with the remaining data:

  1. Everything is passed straight into the history storage as raw data (output 1)
  2. A sliding Window is used to aggregate the average, minimum and maximum values into five minute chunks. This provides a smaller normalized set of values, pre-crunched for quicker analysis. (output 2)

 

Input = IoT Hub - "IoTHubStream"

Event hub name = ThisIsDemo

Event hub policy name = iothubowner

Event hub consumer group = empty (default)

Event serialization format = JSON

Encoding = UTF-8

WITH
[StreamData]
AS (
SELECT
*
FROM
[IoTHubStream]
WHERE
[ObjectType] IS NULL -- Filter out device info and command responses
)
 
SELECT
*
INTO
[Telemetry]
FROM
[StreamData]
SELECT
DeviceId,
AVG (Humidity) AS [AverageHumidity],
MIN(Humidity) AS [MinimumHumidity],
MAX(Humidity) AS [MaxHumidity],
5.0 AS TimeframeMinutes
INTO
[TelemetrySummary]
FROM
[StreamData]
WHERE
[Humidity] IS NOT NULL
GROUP BY
DeviceId,
SlidingWindow (mi, 5)

 

Output 1 : "Telemetry" :

Storage Account = thisisdemo

Container = DeviceTelemetry

Path/pattern = devicetelemetry/{date}/{time}

Date Format = YYYY/MM/DD

Time Format = HH

Serialization = CSV, Delimiter = comma, Encoding = UTF-8

 

Output 2 : "TelemetrySummary"

Storage Account = thisisdemo

Container = DeviceTelemetry

Path/pattern = devicetelemetry-summary/{date}/{time}

Date Format = YYYY/MM/DD

Time Format = HH

Serialization = CSV, Delimiter = comma, Encoding = UTF-8

 

#2 Name: ThisIsDemo-DeviceInfo

The input source is your IoT Hub.

The query simply passes everything that is only device data into a "sub-hub", for other clients like the Web App to consume and monitor.

 

Input = IoT Hub - "DeviceDataStream"

Event hub name = ThisIsDemo

Event hub policy name = iothubowner

Event hub consumer group = empty (default)

Event serialization format = JSON

Encoding = UTF-8

 

SELECT * FROM DeviceDataStream Partition By PartitionId WHERE ObjectType = 'DeviceInfo'

 

Output – Event Hub : DeviceInfoEvents

Event hub name = ThisIsDemo-ehout

Event hub policy name = RootManageSharedAccessKey

Partition key column = PartitionId

Event serialization format = JSON

Encoding = UTF-8

Format = Array

 

#3 Name: ThisIsDemo-Rules

The input source is your IoT Hub.

This query takes any device data that breaks any of the rules defined in the devicerules blob.

Two outputs save the data into blob storage for historical analysis, and the rules Event Hub ThisIsDemo-ehruleout for other clients to listen and act upon.

Stream Analytics supports for refreshing reference data using date/time on the blob file name, described here.  Each time the rules are changed, we save a new blob with a near-future time stamp so the stream job picks it up dynamically.  

 

Input Reference data – blob storage – "DeviceRulesBlob"

Storage Account = thisisdemo

Container = devicerules

Path/pattern = {date}/{time}/devicerules.json

Date Format = YYYY/MM/DD

Serialization = JSON, Encodeing = UTF-8

 

Input = IoT Hub - "IotTelemetryStream"

Event hub name = ThisIsDemo

Event hub policy name = iothubowner

Event hub consumer group = empty (default)

Event serialization format = JSON

Encoding = UTF-8

WITH AlarmsData AS
(
SELECT
Stream.DeviceID, Stream.Temperature as TempReading,
Ref.Temperature as TempThreshold,
Stream.Humidity as HumidityReading,
Ref.Humidity as HumidityThreshold,
Ref.TemperatureRuleOutput,
Ref.HumidityRuleOutput,
Stream.EventEnqueuedUtcTime AS  [Time]
FROM  IoTTelemetryStream Stream
JOIN DeviceRulesBlob Ref
ON Stream.DeviceID = Ref.DeviceID
WHERE
(
Ref.Humidity IS NOT  null AND 
Stream.Humidity > Ref.Humidity) OR
(Ref.Temperature IS NOT  null AND  Stream.Temperature > Ref.Temperature)
)
SELECT * INTO DeviceRulesMonitoring
FROM AlarmsData
SELECT * INTO DeviceRulesHub
FROM AlarmsData

 

Output 1 – Event Hub - "DeviceRulesHub"

Event hub name = ThisIsDemo-ehruleout

Event hub policy name = RootManageSharedAccessKey

Partition key column = PartitionId

Event serialization format = JSON

Encoding = UTF-8

Format = Array

 

Output 2 – Blob Storage - DeviceRulesMonitoring

Storage Account = thisisdemo

Container = rulesoutput

path/pattern = {date}/{time}

Date Format = YYYY/MM/DD

Time Format = HH

Serialization = CSV, Delimiter = comma, Encoding = UTF-8

 

App Service Plan

Name: ThisIsDemo-plan

Pricing Tier : Standard - 2 small

Apps & Slots – 1

 

Lots of useful stats, snapshots, overviews and usage limits

 

Web App

Name: ThisIsDemo 

This is the default web app Microsoft provide to monitor your device data.

App Service plan/pricing tier - ThisIsDemo-plan (Standard: 2 Small)

 

Bing Maps Service

Name: ThisIsDemo-map

Provider = Microsoft.BingMaps

Offer = mapAPIs

Plan = internal1 (FREE)

This service is added to your IoT Solution by IoT Suite to show how well Bing Maps works for such a scenario. You can of course cancel this and replace with whatever you need, like a building floor plan view. However, Bing Maps does provide a lot for free, potentially enough for an internal company system.

"Use up to 10,000 free internal website cumulative transactions of any type, including sessions within any one month period. If you run out, you can upgrade to a paid plan with an increased quota."

"Use up to 500,000 billable internal website cumulative transactions of any type, including sessions within any one month period. Need more than 500,000 transactions per month? Email maplic@microsoft.com for assistance in determining the Bing Maps service and licensing program that’s best for your business."

 

Back to top


Deleting Everything IoT Suite Created

     

When you want to delete the complete solution, and everything that was created by IoT Suite initially and during the example, you simply have to click the delete solution button, as shown below.

Everything that was created for that specific solution, which is everything in the Resource Group and the AAD service that was created for it is deleted.

You can also delete the Resource Group in Azure Portal, but this will not remove the AAD, as explained here.       

 

Back to top


See Also

Another important place to find an extensive amount of Cortana Intelligence Suite related articles is the TechNet Wiki itself. The best entry point is Cortana Intelligence Suite Resources on the TechNet Wiki.