Smartish Edge Camera – Azure IoT Central

This post builds on Smartish Edge Camera – Azure Hub Part 1 using the Azure IoT Hub Device Provisioning Service(DPS) to connect to Azure IoT Central.

The list of object classes is in the YoloCocoP5Model.cs file in the mentalstack/yolov5-net repository.

public override List<YoloLabel> Labels { get; set; } = new List<YoloLabel>()
{
    new YoloLabel { Id = 1, Name = "person" },
    new YoloLabel { Id = 2, Name = "bicycle" },
    new YoloLabel { Id = 3, Name = "car" },
    new YoloLabel { Id = 4, Name = "motorcycle" },
    new YoloLabel { Id = 5, Name = "airplane" },
    new YoloLabel { Id = 6, Name = "bus" },
    new YoloLabel { Id = 7, Name = "train" },
    new YoloLabel { Id = 8, Name = "truck" },
    new YoloLabel { Id = 9, Name = "boat" },
    new YoloLabel { Id = 10, Name = "traffic light" },
    new YoloLabel { Id = 11, Name = "fire hydrant" },
    new YoloLabel { Id = 12, Name = "stop sign" },
    new YoloLabel { Id = 13, Name = "parking meter" },
    new YoloLabel { Id = 14, Name = "bench" },
    new YoloLabel { Id = 15, Name = "bird" },
    new YoloLabel { Id = 16, Name = "cat" },
    new YoloLabel { Id = 17, Name = "dog" },
    new YoloLabel { Id = 18, Name = "horse" },
    new YoloLabel { Id = 19, Name = "sheep" },
    new YoloLabel { Id = 20, Name = "cow" },
    new YoloLabel { Id = 21, Name = "elephant" },
    new YoloLabel { Id = 22, Name = "bear" },
    new YoloLabel { Id = 23, Name = "zebra" },
    new YoloLabel { Id = 24, Name = "giraffe" },
    new YoloLabel { Id = 25, Name = "backpack" },
    new YoloLabel { Id = 26, Name = "umbrella" },
    new YoloLabel { Id = 27, Name = "handbag" },
    new YoloLabel { Id = 28, Name = "tie" },
    new YoloLabel { Id = 29, Name = "suitcase" },
    new YoloLabel { Id = 30, Name = "frisbee" },
    new YoloLabel { Id = 31, Name = "skis" },
    new YoloLabel { Id = 32, Name = "snowboard" },
    new YoloLabel { Id = 33, Name = "sports ball" },
    new YoloLabel { Id = 34, Name = "kite" },
    new YoloLabel { Id = 35, Name = "baseball bat" },
    new YoloLabel { Id = 36, Name = "baseball glove" },
    new YoloLabel { Id = 37, Name = "skateboard" },
    new YoloLabel { Id = 38, Name = "surfboard" },
    new YoloLabel { Id = 39, Name = "tennis racket" },
    new YoloLabel { Id = 40, Name = "bottle" },
    new YoloLabel { Id = 41, Name = "wine glass" },
    new YoloLabel { Id = 42, Name = "cup" },
    new YoloLabel { Id = 43, Name = "fork" },
    new YoloLabel { Id = 44, Name = "knife" },
    new YoloLabel { Id = 45, Name = "spoon" },
    new YoloLabel { Id = 46, Name = "bowl" },
    new YoloLabel { Id = 47, Name = "banana" },
    new YoloLabel { Id = 48, Name = "apple" },
    new YoloLabel { Id = 49, Name = "sandwich" },
    new YoloLabel { Id = 50, Name = "orange" },
    new YoloLabel { Id = 51, Name = "broccoli" },
    new YoloLabel { Id = 52, Name = "carrot" },
    new YoloLabel { Id = 53, Name = "hot dog" },
    new YoloLabel { Id = 54, Name = "pizza" },
    new YoloLabel { Id = 55, Name = "donut" },
    new YoloLabel { Id = 56, Name = "cake" },
    new YoloLabel { Id = 57, Name = "chair" },
    new YoloLabel { Id = 58, Name = "couch" },
    new YoloLabel { Id = 59, Name = "potted plant" },
    new YoloLabel { Id = 60, Name = "bed" },
    new YoloLabel { Id = 61, Name = "dining table" },
    new YoloLabel { Id = 62, Name = "toilet" },
    new YoloLabel { Id = 63, Name = "tv" },
    new YoloLabel { Id = 64, Name = "laptop" },
    new YoloLabel { Id = 65, Name = "mouse" },
    new YoloLabel { Id = 66, Name = "remote" },
    new YoloLabel { Id = 67, Name = "keyboard" },
    new YoloLabel { Id = 68, Name = "cell phone" },
    new YoloLabel { Id = 69, Name = "microwave" },
    new YoloLabel { Id = 70, Name = "oven" },
    new YoloLabel { Id = 71, Name = "toaster" },
    new YoloLabel { Id = 72, Name = "sink" },
    new YoloLabel { Id = 73, Name = "refrigerator" },
    new YoloLabel { Id = 74, Name = "book" },
    new YoloLabel { Id = 75, Name = "clock" },
    new YoloLabel { Id = 76, Name = "vase" },
    new YoloLabel { Id = 77, Name = "scissors" },
    new YoloLabel { Id = 78, Name = "teddy bear" },
    new YoloLabel { Id = 79, Name = "hair drier" },
    new YoloLabel { Id = 80, Name = "toothbrush" }
};

Some of the label choices seem a bit arbitrary(frisbee, surfboard) and American(fire hydrant, baseball bat, baseball glove) It was quite tedious configuring the 80 labels in my Azure IoT Central template.

Azure IoT Central Template with all the YoloV5 labels configured

If there is an object with a label in the PredictionLabelsOfInterest list, a tally of each of the different object classes in the image is sent to an Azure IoT Hub/ Azure IoT Central.

"Application": {
  "DeviceID": "",
  "ImageTimerDue": "0.00:00:15",
  "ImageTimerPeriod": "0.00:00:30",

  "ImageCameraFilepath": "ImageCamera.jpg",
  "ImageMarkedUpFilepath": "ImageMarkedup.jpg",

  "YoloV5ModelPath": "YoloV5/yolov5s.onnx",

  "PredictionScoreThreshold": 0.7,
  "PredictionLabelsOfInterest": [
    "bicycle",
    "person"
  ],
  "PredictionLabelsMinimum": [
    "bicycle",
    "car",
    "person"
  ]
}
My backyard just after the car left (the dry patch in shingle on the right)
Smartish Edge Camera Service console just after car left
Smartish Edge Camera Azure IoT Central graphs showing missing data points

After the You Only Look Once(YOLOV5)+ML.Net+Open Neural Network Exchange(ONNX) plumbing has loaded a timer with a configurable due time and period is started.

private async void ImageUpdateTimerCallback(object state)
{
	DateTime requestAtUtc = DateTime.UtcNow;

	// Just incase - stop code being called while photo already in progress
	if (_cameraBusy)
	{
		return;
	}
	_cameraBusy = true;

	_logger.LogInformation("Image processing start");

	try
	{
#if CAMERA_RASPBERRY_PI
		RaspberryPIImageCapture();
#endif
#if CAMERA_SECURITY
		SecurityCameraImageCapture();
#endif
		List<YoloPrediction> predictions;

		using (Image image = Image.FromFile(_applicationSettings.ImageCameraFilepath))
		{
			_logger.LogTrace("Prediction start");
			predictions = _scorer.Predict(image);
			_logger.LogTrace("Prediction done");

			OutputImageMarkup(image, predictions, _applicationSettings.ImageMarkedUpFilepath);
		}

		if (_logger.IsEnabled(LogLevel.Trace))
		{
			_logger.LogTrace("Predictions {0}", predictions.Select(p => new { p.Label.Name, p.Score }));
		}

		var predictionsValid = predictions.Where(p => p.Score >= _applicationSettings.PredictionScoreThreshold).Select(p => p.Label.Name);

		// Count up the number of each class detected in the image
		var predictionsTally = predictionsValid.GroupBy(p => p)
				.Select(p => new
				{
					Label = p.Key,
					Count = p.Count()
				});

		if (_logger.IsEnabled(LogLevel.Information))
		{
			_logger.LogInformation("Predictions tally before {0}", predictionsTally.ToList());
		}

		// Add in any missing counts the cloudy side is expecting
		if (_applicationSettings.PredictionLabelsMinimum != null)
		{
			foreach( String label in _applicationSettings.PredictionLabelsMinimum)
			{
				if (!predictionsTally.Any(c=>c.Label == label ))
				{
					predictionsTally = predictionsTally.Append(new {Label = label, Count = 0 });
				}
			}
		}

		if (_logger.IsEnabled(LogLevel.Information))
		{
			_logger.LogInformation("Predictions tally after {0}", predictionsTally.ToList());
		}

		if ((_applicationSettings.PredictionLabelsOfInterest == null) || (predictionsValid.Select(c => c).Intersect(_applicationSettings.PredictionLabelsOfInterest, StringComparer.OrdinalIgnoreCase).Any()))
		{
			JObject telemetryDataPoint = new JObject();

			foreach (var predictionTally in predictionsTally)
			{
				telemetryDataPoint.Add(predictionTally.Label, predictionTally.Count);
			}

			using (Message message = new Message(Encoding.ASCII.GetBytes(JsonConvert.SerializeObject(telemetryDataPoint))))
			{
				message.Properties.Add("iothub-creation-time-utc", requestAtUtc.ToString("s", CultureInfo.InvariantCulture));

				await _deviceClient.SendEventAsync(message);
			}
		}
	}
	catch (Exception ex)
	{
		_logger.LogError(ex, "Camera image download, post processing, or telemetry failed");
	}
	finally
	{
		_cameraBusy = false;
	}

	TimeSpan duration = DateTime.UtcNow - requestAtUtc;

	_logger.LogInformation("Image processing done {0:f2} sec", duration.TotalSeconds);
}

Using some Language Integrated Query (LINQ) code any predictions with a score < PredictionScoreThreshold are discarded. A count of the instances of each class is generated with some more LINQ code.

The PredictionLabelsMinimum(optional) is then used to add additional labels with a count of 0 to PredictionsTally so there are no missing datapoints. This is specifically for Azure IoT Central Dashboard so the graph lines are continuous.

Smartish Edge Camera Service console just after put bike in-front of the garage

If any of the list of valid predictions labels is in the PredictionLabelsOfInterest list (if the PredictionLabelsOfInterest is empty any label is a label of interest) the list of prediction class counts is used to populate a Newtonsoft JObject which is serialised to generate a Java Script Object Notation(JSON) Azure IoT Hub message payload.

The “automagic” graph scaling can be sub-optimal

The mentalstack/yolov5-net and NuGet have been incredibly useful and MentalStack team have done a marvelous job building and supporting this project.

The test-rig consisted of a Unv ADZK-10 Security Camera, Power over Ethernet(PoE) and my HP Prodesk 400G4 DM (i7-8700T).

Smartish Edge Camera – Azure IoT Hub

The SmartEdgeCameraAzureIoTService application uses the same You Only Look Once(YOLOV5) + ML.Net + Open Neural Network Exchange(ONNX) plumbing as the SmartEdgeCameraAzureStorageService.

If there is an object with a label in the PredictionLabelsOfInterest list, a tally of each of the different object classes is sent to an Azure IoT Hub.

"Application": {
  "DeviceID": "",
  "ImageTimerDue": "0.00:00:15",
  "ImageTimerPeriod": "0.00:00:30",

  "ImageCameraFilepath": "ImageCamera.jpg",

  "YoloV5ModelPath": "YoloV5/yolov5s.onnx",

  "PredicitionScoreThreshold": 0.7,
  "PredictionLabelsOfInterest": [
    "person"
  ],
}

The Azure IoT hub can configured via a Shared Access Signature(SAS) device policy connection string or the Azure IoT Hub Device Provisioning Service(DPS)

Cars and bicycles in my backyard with no object(s) of interest
SmartEdgeCameraAzureIoTService no object(s) of interest
Cars and bicycles in my backyard with one object of interest
SmartEdgeCameraAzureIoTService one object of interest
Azure IoT Explorer Telemetry with one object of interest

After the You Only Look Once(YOLOV5)+ML.Net+Open Neural Network Exchange(ONNX) plumbing has loaded a timer with a configurable due time and period is started. Using some Language Integrated Query (LINQ) code any predictions with a score < PredictionScoreThreshold are discarded, then the list of predictions is checked to see if there are any in the PredictionLabelsOfInterest. If there are any matching predictions a count of the instances of each class is generated with more LINQ code.

private async void ImageUpdateTimerCallback(object state)
{
	DateTime requestAtUtc = DateTime.UtcNow;

	// Just incase - stop code being called while photo already in progress
	if (_cameraBusy)
	{
		return;
	}
	_cameraBusy = true;

	_logger.LogInformation("Image processing start");

	try
	{
#if CAMERA_RASPBERRY_PI
		RaspberryPIImageCapture();
#endif
#if CAMERA_SECURITY
		SecurityCameraImageCapture();
#endif
		List<YoloPrediction> predictions;

		using (Image image = Image.FromFile(_applicationSettings.ImageCameraFilepath))
		{
			_logger.LogTrace("Prediction start");
			predictions = _scorer.Predict(image);
			_logger.LogTrace("Prediction done");
		}

		if (_logger.IsEnabled(LogLevel.Trace))
		{
			_logger.LogTrace("Predictions {0}", predictions.Select(p => new { p.Label.Name, p.Score }));
		}

		var predictionsOfInterest = predictions.Where(p => p.Score > _applicationSettings.PredicitionScoreThreshold)
										.Select(c => c.Label.Name)
										.Intersect(_applicationSettings.PredictionLabelsOfInterest, StringComparer.OrdinalIgnoreCase);

		if (predictionsOfInterest.Any())
		{
			if (_logger.IsEnabled(LogLevel.Trace))
			{
				_logger.LogTrace("Predictions of interest {0}", predictionsOfInterest.ToList());
			}

			var predictionsTally = predictions.GroupBy(p => p.Label.Name)
									.Select(p => new
									{
										Label = p.Key,
										Count = p.Count()
									});

			if (_logger.IsEnabled(LogLevel.Information))
			{
				_logger.LogInformation("Predictions tally {0}", predictionsTally.ToList());
			}

			JObject telemetryDataPoint = new JObject();

			foreach (var predictionTally in predictionsTally)
			{
				telemetryDataPoint.Add(predictionTally.Label, predictionTally.Count);
			}

			using (Message message = new Message(Encoding.ASCII.GetBytes(JsonConvert.SerializeObject(telemetryDataPoint))))
			{
				message.Properties.Add("iothub-creation-time-utc", requestAtUtc.ToString("s", CultureInfo.InvariantCulture));

				await _deviceClient.SendEventAsync(message);
			}
		}
	}
	catch (Exception ex)
	{
		_logger.LogError(ex, "Camera image download, post processing, telemetry failed");
	}
	finally
	{
		_cameraBusy = false;
	}

	TimeSpan duration = DateTime.UtcNow - requestAtUtc;

	_logger.LogInformation("Image processing done {0:f2} sec", duration.TotalSeconds);
}

The list of prediction class counts is used to populate a Newtonsoft JObject which serialised to generate a Java Script Object Notation(JSON) payload for an Azure IoT Hub message.

The test-rig consisted of a Unv ADZK-10 Security Camera, Power over Ethernet(PoE) and my HP Prodesk 400G4 DM (i7-8700T)

TTI V3 Connector Azure IoT Central Device Provisioning Service(DPS) support

The TTI Connector supports the Azure IoT Hub Device Provisioning Service(DPS) which is required (it is possible to provision individual devices but this intended for small deployments or testing) for Azure IoT Central applications. The TTI Connector implementation also supports Azure IoT Central Digital Twin Definition Language (DTDL V2) for “automagic” device provisioning.

The first step was to configure and Azure IoT Central enrollment group (ensure “Automatically connect devices in this group” is on for “zero touch” provisioning) and copy the IDScope and Group Enrollment key to the TTI Connector configuration

RAK3172 Enrollment Group creation
Azure IoT Hub Device Provisioning Service configuration

I then created an Azure IoT Central template for my RAK3172 breakout board based.Net Core powered test device.

{
    "@id": "dtmi:ttnv3connectorclient:RASK3172Breakout1c7;1",
    "@type": "Interface",
    "contents": [
        {
            "@id": "dtmi:ttnv3connectorclient:RASK3172Breakout1c7:temperature_0;1",
            "@type": [
                "Telemetry",
                "Temperature"
            ],
            "displayName": {
                "en": "Temperature"
            },
            "name": "temperature_0",
            "schema": "double",
            "unit": "degreeCelsius"
        },
        {
            "@id": "dtmi:ttnv3connectorclient:RASK3172Breakout1c7:relative_humidity_0;1",
            "@type": [
                "Telemetry",
                "RelativeHumidity"
            ],
            "displayName": {
                "en": "Humidity"
            },
            "name": "relative_humidity_0",
            "schema": "double",
            "unit": "percent"
        },
        {
            "@id": "dtmi:ttnv3connectorclient:RASK3172Breakout1c7:value_0;1",
            "@type": "Command",
            "displayName": {
                "en": "Temperature OOB alert minimum"
            },
            "name": "value_0",
            "request": {
                "@type": "CommandPayload",
                "displayName": {
                    "en": "Minimum"
                },
                "name": "value_0",
                "schema": "double"
            },
            "durable": true
        },
        {
            "@id": "dtmi:ttnv3connectorclient:RASK3172Breakout1c7:value_1;1",
            "@type": "Command",
            "displayName": {
                "en": "Temperature OOB alert maximum"
            },
            "name": "value_1",
            "request": {
                "@type": "CommandPayload",
                "displayName": {
                    "en": "Maximum"
                },
                "name": "value_1",
                "schema": "double"
            },
            "durable": true
        },
        {
            "@id": "dtmi:ttnv3connectorclient:RASK3172Breakout1c7:TemperatureOOBAlertMinimumAndMaximum;1",
            "@type": "Command",
            "displayName": {
                "en": "Temperature OOB alert minimum and maximum"
            },
            "name": "TemperatureOOBAlertMinimumAndMaximum",
            "request": {
                "@type": "CommandPayload",
                "displayName": {
                    "en": "Alert Temperature"
                },
                "name": "AlertTemperature",
                "schema": {
                    "@type": "Object",
                    "displayName": {
                        "en": "Object"
                    },
                    "fields": [
                        {
                            "displayName": {
                                "en": "minimum"
                            },
                            "name": "value_0",
                            "schema": "double"
                        },
                        {
                            "displayName": {
                                "en": "maximum"
                            },
                            "name": "value_1",
                            "schema": "double"
                        }
                    ]
                }
            },
            "durable": true
        },
        {
            "@id": "dtmi:ttnv3connectorclient:RASK3172Breakout1c7:value_2;1",
            "@type": "Command",
            "displayName": {
                "en": "Fan"
            },
            "name": "value_2",
            "request": {
                "@type": "CommandPayload",
                "displayName": {
                    "en": "On"
                },
                "name": "value_3",
                "schema": {
                    "@type": "Enum",
                    "displayName": {
                        "en": "Enum"
                    },
                    "enumValues": [
                        {
                            "displayName": {
                                "en": "On"
                            },
                            "enumValue": 1,
                            "name": "On"
                        },
                        {
                            "displayName": {
                                "en": "Off"
                            },
                            "enumValue": 0,
                            "name": "Off"
                        }
                    ],
                    "valueSchema": "integer"
                }
            },
            "durable": true
        },
        {
            "@id": "dtmi:ttnv3connectorclient:RASK3172Breakout1c7:LightsGoOn;1",
            "@type": "Command",
            "displayName": {
                "en": "LightsGoOn"
            },
            "name": "LightsGoOn",
            "durable": true
        },
        {
            "@id": "dtmi:ttnv3connectorclient:RASK3172Breakout1c7:LightsGoOff;1",
            "@type": "Command",
            "displayName": {
                "en": "LightsGoOff"
            },
            "name": "LightsGoOff",
            "durable": true
        }
    ],
    "displayName": {
        "en": "RASK3172 Breakout"
    },
    "@context": [
        "dtmi:iotcentral:context;2",
        "dtmi:dtdl:context;2"
    ]
}

The Device Template @Id can also be set for a TTI application using an optional dtdlmodelid which is specified the the TTI Connector configuration.

Azure Smartish Edge Camera – The basics

This project builds on my ML.Net YoloV5 + Camera + GPIO on ARM64 Raspberry PI with the addition of basic support for Azure IoT Hubs, the Azure IoT Hub Device Provisioning Service(DPS), and Azure IoT Central.

My backyard test-rig has consists of a Unv ADZK-10 Security Camera, Power over Ethernet(PoE) module, and an ASUS PE100A.

Backyard test-rig

The application can be compiled with support for Azure IoT Connection strings or the Device Provisioning Service(DPS). The appsetings.json file has configuration options for Azure IoT Hub connection string or DPS Global Device Endpoint+ScopeID+Group Enrollment key.

{
  "ApplicationSettings": {
    "DeviceId": "NotTheEdgeCamera",

    "ImageTimerDue": "0.00:00:15",
    "ImageTimerPeriod": "0.00:00:30",

    "CameraUrl": "http://10.0.0.55:85/images/snapshot.jpg",
    "CameraUserName": ",,,",
    "CameraUserPassword": "...",

    "ButtonPinNumer": 6,
    "LedPinNumer": 5,

    "InputImageFilenameLocal": "InputLatest.jpg",
    "OutputImageFilenameLocal": "OutputLatest.jpg",

    "ProcessWaitForExit": 10000,

    "YoloV5ModelPath": "Assets/YoloV5/yolov5s.onnx",

    "PredicitionScoreThreshold": 0.5,

    "AzureIoTHubConnectionString": "...",

    "GlobalDeviceEndpoint": "global.azure-devices-provisioning.net",
    "AzureIoTHubDpsIDScope": "...",
    "AzureIoTHubDpsGroupEnrollmentKey": "..."
  }
}

After the You Only Look Once(YOLOV5)+ML.Net+Open Neural Network Exchange(ONNX) plumbing has loaded a timer with a configurable due time and period is started.

private static async void ImageUpdateTimerCallback(object state)
{
	DateTime requestAtUtc = DateTime.UtcNow;

	// Just incase - stop code being called while photo already in progress
	if (_cameraBusy)
	{
		return;
	}
	_cameraBusy = true;

	Console.WriteLine($"{DateTime.UtcNow:yy-MM-dd HH:mm:ss} Image processing start");

	try
	{
#if SECURITY_CAMERA
		Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss:fff} Security Camera Image download start");
		SecurityCameraImageCapture();
		Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss:fff} Security Camera Image download done");
#endif

#if RASPBERRY_PI_CAMERA
		Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss:fff} Raspberry PI Image capture start");
		RaspberryPICameraImageCapture();
		Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss:fff} Raspberry PI Image capture done");
#endif

		List<YoloPrediction> predictions;

		// Process the image on local file system
		using (Image image = Image.FromFile(_applicationSettings.InputImageFilenameLocal))
		{
			Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss:fff} YoloV5 inferencing start");
			predictions = _scorer.Predict(image);
			Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss:fff} YoloV5 inferencing done");

#if OUTPUT_IMAGE_MARKUP
			using (Graphics graphics = Graphics.FromImage(image))
			{
				Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss:fff} Image markup start");

				foreach (var prediction in predictions) // iterate predictions to draw results
				{
					double score = Math.Round(prediction.Score, 2);

					graphics.DrawRectangles(new Pen(prediction.Label.Color, 1), new[] { prediction.Rectangle });

					var (x, y) = (prediction.Rectangle.X - 3, prediction.Rectangle.Y - 23);

					graphics.DrawString($"{prediction.Label.Name} ({score})", new Font("Consolas", 16, GraphicsUnit.Pixel), new SolidBrush(prediction.Label.Color), new PointF(x, y));
				}

				image.Save(_applicationSettings.OutputImageFilenameLocal);

				Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss:fff} Image markup done");
			}
#endif
		}

#if AZURE_IOT_HUB_CONNECTION || AZURE_IOT_HUB_DPS_CONNECTION
		await AzureIoTHubTelemetry(requestAtUtc, predictions);
#endif
	}
	catch (Exception ex)
	{
		Console.WriteLine($"{DateTime.UtcNow:yy-MM-dd HH:mm:ss} Camera image download, post procesing, image upload, or telemetry failed {ex.Message}");
	}
	finally
	{
		_cameraBusy = false;
	}

	TimeSpan duration = DateTime.UtcNow - requestAtUtc;

	Console.WriteLine($"{DateTime.UtcNow:yy-MM-dd HH:mm:ss} Image processing done {duration.TotalSeconds:f2} sec");
	Console.WriteLine();
}

In the ImageUpdateTimerCallback method a camera image is captured (Raspberry Pi Camera Module 2 or Unv ADZK-10 Security Camera) and written to the local file system.

SSH Connection to Azure PE100 running Smartish Camera application

The YoloV5 model ML.Net support library then loads the image and processes the prediction output (can be inspected with Netron) generating list of objects that have been detected, their Minimum Bounding Rectangle(MBR) and class.

public static async Task AzureIoTHubTelemetry(DateTime requestAtUtc, List<YoloPrediction> predictions)
{
	JObject telemetryDataPoint = new JObject();

	foreach (var predictionTally in predictions.Where(p => p.Score >= _applicationSettings.PredicitionScoreThreshold).GroupBy(p => p.Label.Name)
					.Select(p => new
					{
						Label = p.Key,
						Count = p.Count()
					}))
	{
		Console.WriteLine("  {0} {1}", predictionTally.Label, predictionTally.Count);

		telemetryDataPoint.Add(predictionTally.Label, predictionTally.Count);
	}

	try
	{
		using (Message message = new Message(Encoding.ASCII.GetBytes(JsonConvert.SerializeObject(telemetryDataPoint))))
		{
			message.Properties.Add("iothub-creation-time-utc", requestAtUtc.ToString("s", CultureInfo.InvariantCulture));

			Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss} AzureIoTHubClient SendEventAsync prediction information start");
			await _deviceClient.SendEventAsync(message);
			Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss} AzureIoTHubClient SendEventAsync prediction information finish");
		}
	}
	catch (Exception ex)
	{
		Console.WriteLine($"{DateTime.UtcNow:yy-MM-dd HH:mm:ss} AzureIoTHubClient SendEventAsync cow counting failed {ex.Message}");
	}
}

The list of predictions is post processed with a Language Integrated Query(LINQ) which filters out predictions with a score below a configurable threshold and returns a count of each class.

My backyard from the deck

The aggregated YoloV5 prediction results are then uploaded to an Azure IoT Hub or Azure IoT Central

Azure IoT Explorer Displaying message payloads from the Smartish Edge Camera
Azure IoT Central displaying message payloads from the Smartish Edge Camera

TTI V3 Connector Device Provisioning Service(DPS) support

The previous versions of my Things Network Industries(TTI) and The Things Network(TTN) connectors supported the Azure IoT Hub Device Provisioning Service(DPS) with Symmetric Key Attestation(SAS) to “automagically” setup the LoRaWAN devices in a TTI Application.(See my V2 Gateway DPS setup post for more detail).

Azure Device Provisioning Service configuring Azure IoT Hubs

I used an “evenly weighted distribution” to spread the devices across five Azure IoT Hubs.

Azure IoT Hub no registered devices

In the Azure Portal I configured the DPS ID Scope (AzureSettings:DeviceProvisioningServiceSettings:IdScope) and the Group Enrollment Key(AzureSettings:DeviceProvisioningServiceSettings:GroupEnrollmentKey) then saved the configuration which restarted the AppService.

Azure Portal AppService configration

The first time a device sent an uplink message the cache query fails and the RegisterAsync method of the ProvisioningDeviceClient is called to get a device connection string.

	logger.LogInformation("Uplink-ApplicationID:{0} DeviceID:{1} Port:{2} Payload Raw:{3}", applicationId, deviceId, port, payload.UplinkMessage.PayloadRaw);

	if (!_DeviceClients.TryGetValue(deviceId, out DeviceClient deviceClient))
	{
		logger.LogInformation("Uplink-Unknown device for ApplicationID:{0} DeviceID:{1}", applicationId, deviceId);

		// Check that only one of Azure Connection string or DPS is configured
		if (string.IsNullOrEmpty(_azureSettings.IoTHubConnectionString) && (_azureSettings.DeviceProvisioningServiceSettings == null))
		{
			logger.LogError("Uplink-Neither Azure IoT Hub connection string or Device Provisioning Service configured");

			return req.CreateResponse(HttpStatusCode.UnprocessableEntity);
		}

		// Check that only one of Azure Connection string or DPS is configured
		if (!string.IsNullOrEmpty(_azureSettings.IoTHubConnectionString) && (_azureSettings.DeviceProvisioningServiceSettings != null))
		{
			logger.LogError("Uplink-Both Azure IoT Hub connection string and Device Provisioning Service configured");

			return req.CreateResponse(HttpStatusCode.UnprocessableEntity);
		}

		// User Azure IoT Connection string if configured and Device Provisioning Service isn't
		if (!string.IsNullOrEmpty(_azureSettings.IoTHubConnectionString))
		{
			deviceClient = DeviceClient.CreateFromConnectionString(_azureSettings.IoTHubConnectionString, deviceId, transportSettings);

			try
			{
				await deviceClient.OpenAsync();
			}
			catch (DeviceNotFoundException)
			{
				logger.LogWarning("Uplink-Unknown DeviceID:{0}", deviceId);

				return req.CreateResponse(HttpStatusCode.NotFound);
			}
		}

		// Azure IoT Hub Device provisioning service if configured
		if (_azureSettings.DeviceProvisioningServiceSettings != null) 
		{
			string deviceKey;

			if ( string.IsNullOrEmpty(_azureSettings.DeviceProvisioningServiceSettings.IdScope) || string.IsNullOrEmpty(_azureSettings.DeviceProvisioningServiceSettings.GroupEnrollmentKey))
			{
				logger.LogError("Uplink-Device Provisioning Service requires ID Scope and Group Enrollment Key configured");

				return req.CreateResponse(HttpStatusCode.UnprocessableEntity);
			}

			using (var hmac = new HMACSHA256(Convert.FromBase64String(_azureSettings.DeviceProvisioningServiceSettings.GroupEnrollmentKey)))
			{
				deviceKey = Convert.ToBase64String(hmac.ComputeHash(Encoding.UTF8.GetBytes(deviceId)));
			}

			using (var securityProvider = new SecurityProviderSymmetricKey(deviceId, deviceKey, null))
			{
				using (var transport = new ProvisioningTransportHandlerAmqp(TransportFallbackType.TcpOnly))
				{
					ProvisioningDeviceClient provClient = ProvisioningDeviceClient.Create(
						Constants.AzureDpsGlobalDeviceEndpoint,
						_azureSettings.DeviceProvisioningServiceSettings.IdScope,
						securityProvider,
						transport);

					DeviceRegistrationResult result = await provClient.RegisterAsync();

					if (result.Status != ProvisioningRegistrationStatusType.Assigned)
					{
						_logger.LogError("Config-DeviceID:{0} Status:{1} RegisterAsync failed ", deviceId, result.Status);

						return req.CreateResponse(HttpStatusCode.FailedDependency);
					}

					IAuthenticationMethod authentication = new DeviceAuthenticationWithRegistrySymmetricKey(result.DeviceId, (securityProvider as SecurityProviderSymmetricKey).GetPrimaryKey());

					deviceClient = DeviceClient.Create(result.AssignedHub, authentication, transportSettings);

					await deviceClient.OpenAsync();
				}
			}
		}

		if (!_DeviceClients.TryAdd(deviceId, deviceClient))
		{
			logger.LogWarning("Uplink-TryAdd failed for ApplicationID:{0} DeviceID:{1}", applicationId, deviceId);

			return req.CreateResponse(HttpStatusCode.Conflict);
		}

		Models.AzureIoTHubReceiveMessageHandlerContext context = new Models.AzureIoTHubReceiveMessageHandlerContext()
		{
			DeviceId = deviceId,
			ApplicationId = applicationId,
			WebhookId = _theThingsIndustriesSettings.WebhookId,
			WebhookBaseURL = _theThingsIndustriesSettings.WebhookBaseURL,
			ApiKey = _theThingsIndustriesSettings.ApiKey
		};

		await deviceClient.SetReceiveMessageHandlerAsync(AzureIoTHubClientReceiveMessageHandler, context);

		await deviceClient.SetMethodDefaultHandlerAsync(AzureIoTHubClientDefaultMethodHandler, context);
	}

	JObject telemetryEvent = new JObject
	{
		{ "ApplicationID", applicationId },
		{ "DeviceID", deviceId },
		{ "Port", port },
		{ "Simulated", payload.Simulated },
		{ "ReceivedAtUtc", payload.UplinkMessage.ReceivedAtUtc.ToString("s", CultureInfo.InvariantCulture) },
		{ "PayloadRaw", payload.UplinkMessage.PayloadRaw }
	};

	// If the payload has been decoded by payload formatter, put it in the message body.
	if (payload.UplinkMessage.PayloadDecoded != null)
	{
		telemetryEvent.Add("PayloadDecoded", payload.UplinkMessage.PayloadDecoded);
	}

	// Send the message to Azure IoT Hub
	using (Message ioTHubmessage = new Message(Encoding.ASCII.GetBytes(JsonConvert.SerializeObject(telemetryEvent))))
	{
		// Ensure the displayed time is the acquired time rather than the uploaded time. 
		ioTHubmessage.Properties.Add("iothub-creation-time-utc", payload.UplinkMessage.ReceivedAtUtc.ToString("s", CultureInfo.InvariantCulture));
		ioTHubmessage.Properties.Add("ApplicationId", applicationId);
		ioTHubmessage.Properties.Add("DeviceEUI", payload.EndDeviceIds.DeviceEui);
		ioTHubmessage.Properties.Add("DeviceId", deviceId);
		ioTHubmessage.Properties.Add("port", port.ToString());
		ioTHubmessage.Properties.Add("Simulated", payload.Simulated.ToString());

		await deviceClient.SendEventAsync(ioTHubmessage);

		logger.LogInformation("Uplink-DeviceID:{0} SendEventAsync success", payload.EndDeviceIds.DeviceId);
	}
}
catch (Exception ex)
{
	logger.LogError(ex, "Uplink-Message processing failed");

	return req.CreateResponse(HttpStatusCode.InternalServerError);
}

I used Telerik Fiddler and some sample payloads copied from my Azure Storage Queue sample to simulate many devices and the registrations were spread across my five Azure IoT Hubs.

DPS Device Registrations tab showing distribution of LoRaWAN Devices

I need to review the HTTP Error codes returned for different errors and ensure failures are handled robustly.

TTN V3 Connector Revisited

Earlier in the year I built Things Network(TTN) V2 and V3 connectors and after using these in production applications I have learnt a lot about what I had got wrong, less wrong and what I had got right.

Using a TTN V3 MQTT Application integration wasn’t a great idea. The management of state was very complex. The storage of application keys in a app.settings file made configuration easy but was bad for security.

The use of Azure Key Vault in the TTNV2 connector was a good approach, but the process of creation and updating of the settings needs to be easier.

Using TTN device registry as the “single source of truth” was a good decision as managing the amount of LoRaWAN network, application and device specific configuration in an Azure IoT Hub would be non-trivial.

Using a Webhooks Application Integration like the TTNV2 connector is my preferred approach.

The TTNV2 Connector’s use of Azure Storage Queues was a good idea as they it provide an elastic buffer between the different parts of the application.

The use of Azure Functions to securely ingest webhook calls and write them to Azure Storage Queues with output bindingts should simplify configuration and deployment. The use of Azure Storage Queue input bindings to process messages is the preferred approach.

The TTN V3 processing of JSON uplink messages into a structure that Azure IoT Central could ingest is a required feature

The TTN V2 and V3 support for the Azure Device Provisioning Service(DPS) is a required feature (mandated by Azure IoT Central). The TTN V3 connector support for DTDLV2 is a desirable feature. The DPS implementation worked with Azure IoT Central but I was unable to get the DeviceClient based version working.

Using DPS to pre-provision devices in Azure IoT Hubs and Azure IoT Central by using the TTN Application Registry API then enumerating the TTN applications, then devices needs to be revisited as it was initially slow then became quite complex.

The support for Azure IoT Hub connection strings was a useful feature, but added some complexity. This plus basic Azure IoT Hub DPS support(No Azure IoT Central support) could be implemented in a standalone application which connects via Azure Storage Queue messages.

The processing of Azure IoT Central Basic, and Request commands then translating the payloads so they work with TTN V3 is a required feature. The management of Azure IoT Hub command delivery confirmations (abandon, complete and Reject) is a required feature.

I’m considering building a new TTN V3 connector but is it worth the effort as TTN has one now?

Device Provisioning Service(DPS) JsonData

While building my The Things Industries(TTI) V3 connector which uses the Azure Device Provisioning Service(DPS) the way pretty much all of the samples formatted the JsonData property of the ProvisioningRegistrationAdditionalData (part of Plug n Play provisioning) by manually constructing a JSON object which bugged me.

ProvisioningRegistrationAdditionalData provisioningRegistrationAdditionalData = new ProvisioningRegistrationAdditionalData()
{
   JsonData = $"{{\"modelId\": \"{modelId}\"}}"
};

result = await provClient.RegisterAsync(provisioningRegistrationAdditionalData);

I remembered seeing a sample where the DTDLV2 methodId was formatted by a library function and after a surprising amount of searching I found what I was looking for in Azure-Samples repository.

The code for the CreateDpsPayload method

// Copyright (c) Microsoft. All rights reserved.
// Licensed under the MIT license. See LICENSE file in the project root for full license information.

using Microsoft.Azure.Devices.Provisioning.Client.Extensions;

namespace Microsoft.Azure.Devices.Provisioning.Client.PlugAndPlay
{
    /// <summary>
    /// A helper class for formatting the DPS device registration payload, per plug and play convention.
    /// </summary>
    public static class PnpConvention
    {
        /// <summary>
        /// Create the DPS payload to provision a device as plug and play.
        /// </summary>
        /// <remarks>
        /// For more information on device provisioning service and plug and play compatibility,
        /// and PnP device certification, see <see href="https://docs.microsoft.com/en-us/azure/iot-pnp/howto-certify-device"/>.
        /// The DPS payload should be in the format:
        /// <code>
        /// {
        ///   "modelId": "dtmi:com:example:modelName;1"
        /// }
        /// </code>
        /// For information on DTDL, see <see href="https://github.com/Azure/opendigitaltwins-dtdl/blob/master/DTDL/v2/dtdlv2.md"/>
        /// </remarks>
        /// <param name="modelId">The Id of the model the device adheres to for properties, telemetry, and commands.</param>
        /// <returns>The DPS payload to provision a device as plug and play.</returns>
        public static string CreateDpsPayload(string modelId)
        {
            modelId.ThrowIfNullOrWhiteSpace(nameof(modelId));
            return $"{{\"modelId\":\"{modelId}\"}}";
        }
    }
}

With a couple of changes my code now uses the CreateDpsPayload method

using Microsoft.Azure.Devices.Provisioning.Client.PlugAndPlay;

...

using (var securityProvider = new SecurityProviderSymmetricKey(deviceId, deviceKey, null))
{
   using (var transport = new ProvisioningTransportHandlerAmqp(TransportFallbackType.TcpOnly))
   {
      ProvisioningDeviceClient provClient = ProvisioningDeviceClient.Create(
         Constants.AzureDpsGlobalDeviceEndpoint,
         deviceProvisiongServiceSettings.IdScope,
         securityProvider,
         transport);

      DeviceRegistrationResult result;

      if (!string.IsNullOrEmpty(modelId))
      {
         ProvisioningRegistrationAdditionalData provisioningRegistrationAdditionalData = new ProvisioningRegistrationAdditionalData()
         {
               JsonData = PnpConvention.CreateDpsPayload(modelId)
         };

         result = await provClient.RegisterAsync(provisioningRegistrationAdditionalData, stoppingToken);
      }
      else
      {
         result = await provClient.RegisterAsync(stoppingToken);
      }

      if (result.Status != ProvisioningRegistrationStatusType.Assigned)
      {
         _logger.LogError("Config-DeviceID:{0} Status:{1} RegisterAsync failed ", deviceId, result.Status);

         return false;
      }

      IAuthenticationMethod authentication = new DeviceAuthenticationWithRegistrySymmetricKey(result.DeviceId, (securityProvider as SecurityProviderSymmetricKey).GetPrimaryKey());

      deviceClient = DeviceClient.Create(result.AssignedHub, authentication, transportSettings);
   }
}

TTI V3 Gateway Device Provisioning Service(DPS) Concurrent Requests

While debugging The Things Industries(TTI) V3 connector on my desktop I had noticed that using an Azure IoT Hub device connection string was quite a bit faster than using the Azure Device Provisioning Service(DPS). The Azure Webjob connector was executing the requests sequentially which made the duration of the DPS call even more apparent.

To reduce the impact of the RegisterAsync call duration this Proof of Concept(PoC) code uses the System.Tasks.Threading library to execute each request in its own thread and then wait for all the requests to finish.

try
{
   int devicePage = 1;
   V3EndDevices endDevices = await endDeviceRegistryClient.ListAsync(
      applicationSetting.Key,
      field_mask_paths: Constants.DevicefieldMaskPaths,
      page: devicePage,
      limit: _programSettings.TheThingsIndustries.DevicePageSize,
      cancellationToken: stoppingToken);

   while ((endDevices != null) && (endDevices.End_devices != null)) // If no devices returns null rather than empty list
   {
      List<Task<bool>> tasks = new List<Task<bool>>();

      _logger.LogInformation("Config-ApplicationID:{0} start", applicationSetting.Key);

      foreach (V3EndDevice device in endDevices.End_devices)
      {
         if (DeviceAzureEnabled(device))
         {
            _logger.LogInformation("Config-ApplicationID:{0} DeviceID:{1} Device EUI:{2}", device.Ids.Application_ids.Application_id, device.Ids.Device_id, BitConverter.ToString(device.Ids.Dev_eui));

            tasks.Add(DeviceRegistration(device.Ids.Application_ids.Application_id,
                                       device.Ids.Device_id,
                                       _programSettings.ResolveDeviceModelId(device.Ids.Application_ids.Application_id, device.Attributes),
                                       stoppingToken));
         }
      }

      _logger.LogInformation("Config-ApplicationID:{0} Page:{1} processing start", applicationSetting.Key, devicePage);

      Task.WaitAll(tasks.ToArray(),stoppingToken);

      _logger.LogInformation("Config-ApplicationID:{0} Page:{1} processing finish", applicationSetting.Key, devicePage);

      endDevices = await endDeviceRegistryClient.ListAsync(
         applicationSetting.Key,
         field_mask_paths: Constants.DevicefieldMaskPaths,
         page: devicePage += 1,
         limit: _programSettings.TheThingsIndustries.DevicePageSize,
         cancellationToken: stoppingToken);
   }
   _logger.LogInformation("Config-ApplicationID:{0} finish", applicationSetting.Key);
}
catch (ApiException ex)
{
   _logger.LogError("Config-Application configuration API error:{0}", ex.StatusCode);
}

The connector application paginates the retrieval of device configuration from TTI API and a Task is created for each device returned in a page. In the Application Insights Trace logging the duration of a single page of device registrations was approximately the duration of the longest call.

There will be a tradeoff between device page size (resource utilisation by many threads) and startup duration (to many sequential page operations) which will need to be explored.

TTI V3 Gateway Device Provisioning Service(DPS) Performance

My The Things Industries(TTI) V3 connector is an Identity Translation Cloud Gateway, it maps LoRaWAN devices to Azure IoT Hub devices. The connector creates a DeviceClient for each TTI LoRaWAN device and can use an Azure Device Connection string or the Azure Device Provisioning Service(DPS).

While debugging the connector on my desktop I had noticed that using a connection string was quite a bit faster than using DPS and I had assumed this was just happenstance. While doing some testing in the Azure North Europe data-center (Closer to TTI European servers) I grabbed some screen shots of the trace messages in Azure Application Insights as the TTI Connector Application was starting.

I only have six LoRaWAN devices configured in my TTI dev instance, but I repeated each test several times and the results were consistent so the request durations are reasonable. My TTI Connector application, IoT Hub, DPS and Application insights instances are all in the same Azure Region and Azure Resource Group so networking overheads shouldn’t be significant.

Azure IoT Hub Connection device connection string

Using an Azure IoT Hub Device Shared Access policy connection string establishing a connection took less than a second.

My Azure DPS Instance

Using my own DPS instance to provide the connection string and then establishing a connection took between 3 and 7 seconds.

Azure IoT Central DPS

For my Azure IoT Central instance getting a connection string and establishing a connection took between 4 and 7 seconds.

The Azure DPS client code was copied from one of the sample applications so I have assumed it is “correct”.

using (var transport = new ProvisioningTransportHandlerAmqp(TransportFallbackType.TcpOnly))
{
	ProvisioningDeviceClient provClient = ProvisioningDeviceClient.Create( 
		Constants.AzureDpsGlobalDeviceEndpoint,
		deviceProvisiongServiceSettings.IdScope,
		securityProvider,
		transport);

	DeviceRegistrationResult result;

	if (!string.IsNullOrEmpty(modelId))
	{
		ProvisioningRegistrationAdditionalData provisioningRegistrationAdditionalData = new ProvisioningRegistrationAdditionalData()
		{
			JsonData = $"{{"modelId": "{modelId}"}}"
		};

		result = await provClient.RegisterAsync(provisioningRegistrationAdditionalData, stoppingToken);
	}
	else
    {
		result = await provClient.RegisterAsync(stoppingToken);
	}

	if (result.Status != ProvisioningRegistrationStatusType.Assigned)
	{
		_logger.LogError("Config-DeviceID:{0} Status:{1} RegisterAsync failed ", deviceId, result.Status);

		return false;
	}

	IAuthenticationMethod authentication = new DeviceAuthenticationWithRegistrySymmetricKey(result.DeviceId, (securityProvider as SecurityProviderSymmetricKey).GetPrimaryKey());

	deviceClient = DeviceClient.Create(result.AssignedHub, authentication, transportSettings);
}

I need to investigate why getting a connection string from the DPS then connecting take significantly longer (I appreciate that “behind the scenes” service calls maybe required). This wouldn’t be an issue for individual devices connecting from different locations but for my Identity Translation Cloud gateway which currently open connections sequentially this could be a problem when there are a large number of devices.

If the individual requests duration can’t be reduced (using connection pooling etc.) I may have to spin up multiple threads so multiple devices can be connecting concurrently.

TTN V3 Gateway Downlink Broken

While adding Azure Device Provisioning Service (DPS) support to my The Things Industries(TTI)/The Things Network(TTN) Azure IoT Hub/Azure IoT Central Connector I broke Cloud to Device(C2D)/Downlink messaging. I had copied the Advanced Message Queuing Protocol(AMQP) connection pooling configuration code from my The Things Network Integration assuming it worked.

return DeviceClient.CreateFromConnectionString(connectionString, deviceId,
	new ITransportSettings[]
	{
		new AmqpTransportSettings(TransportType.Amqp_Tcp_Only)
		{
			PrefetchCount = 0,
			AmqpConnectionPoolSettings = new AmqpConnectionPoolSettings()
			{
				Pooling = true,
			}
		}
	});

I hadn’t noticed this issue in my Azure IoT The Things Network Integration because I hadn’t built support for C2D messaging. After some trial and error I figured out the issue was the PrefetchCount initialisation.

return DeviceClient.CreateFromConnectionString(connectionString, deviceId,
	new ITransportSettings[]
	{
		new AmqpTransportSettings(TransportType.Amqp_Tcp_Only)
		{
			AmqpConnectionPoolSettings = new AmqpConnectionPoolSettings()
			{
				Pooling = true,
			}
		}
	});

From the Azure Service Bus (I couldn’t find any specifically Azure IoT Hub ) documentation

Even though the Service Bus APIs do not directly expose such an option today, a lower-level AMQP protocol client can use the link-credit model to turn the “pull-style” interaction of issuing one unit of credit for each receive request into a “push-style” model by issuing a large number of link credits and then receive messages as they become available without any further interaction. Push is supported through the MessagingFactory.PrefetchCount or MessageReceiver.PrefetchCount property settings. When they are non-zero, the AMQP client uses it as the link credit.

n this context, it’s important to understand that the clock for the expiration of the lock on the message inside the entity starts when the message is taken from the entity, not when the message is put on the wire. Whenever the client indicates readiness to receive messages by issuing link credit, it is therefore expected to be actively pulling messages across the network and be ready to handle them. Otherwise the message lock may have expired before the message is even delivered. The use of link-credit flow control should directly reflect the immediate readiness to deal with available messages dispatched to the receiver.

In the Azure IoT Hub SDK the prefetch count is set to 50 (around line 57) and throws an exception if less that zero (around line 90) and there is some information about tuning the prefetch value for Azure Service Bus.

The best explanation I count find was Github issue which was a query “What exactly does the PrefetchCount property control?”

“You are correct, the pre-fetch count is used to set the link credit over AMQP. What this signifies is the max. no. of messages that can be “in-flight” from the service to the client, at any given time. (This value defaults to 50 for the IoT Hub .NET client).
The client specifies its link-credit, that the service must respect. In simplest terms, any time the service sends a message to the client, it decrements the link credit, and will continue sending messages until linkCredit > 0. Once the client acknowledges the message, it will increment the link credit.”

In summary if Prefetch count is set to zero on startup in my application no messages will be sent to the client….