I'm looking for a way to scale a Azure Function properly but I have a problem.
I have a set of IoT devices that send data to Azure by HTTP (For this there is a set of Azure Functions that scale automatically)
But now there is a new IoT device that sends the data via TCP/IP Persistent connections and as far as I know this ins't supported by Azure Functions.
Anyone has some ideia on how to implement this?
One option I thought was to have a VM treating the TCP connections and sending the data to a queue so that a function can pull from the queue in a scalable manner.
Thanks in advance, cheers.
If your device only communicates via a persistent TCP/IP connection, then you'll need to implement an IoT Gateway. An IoT Gateway basically an app that either runs on-premises or in the cloud, it communicates with 1 or more IoT devices, then the Gateway can connect to and communicate with a service like Azure IoT Hub. From there, you can get your events sent up to Azure IoT Hub, then process those events in the cloud by using something like Azure Stream Analytics to send then to Event Hubs, or Service Bus Queue, as example. Then once in a Queue you can write an Azure Function to be triggered to handle those events as they come int.
There is no way to open a persistent TCP/IP connection to an Azure Function. Azure Functions are Serverless compute, and are built to be Event-Driven in nature and short running. It is recommended that an Azure Function runs for for a maximum duration of less than 5 minutes per execution.
I think your best bet is decoupling from the device protocols; functions can bind to Event Hubs, and you can put your proxies in containers (pretty easy to host and scale those on App Svcs or AKS).
This way the brittle part (the proxies) are easy to test.
Something like fanout.io might work depending on the protocol. If it's MQTT, going Azure IoT Hub may make sense.
Related
I have a local machine that reads RabbitMQ queue messages.
I wish to move it to cloud. Which Azure service can be used in this case?
I went through event hubs, but I am not sure, if it would read messages from rabbitMQ continuously.
Any suggestions for the service that should be put to use.
You should take a look at Azure Service Bus. It has got FIFO queues as well as publish/subscribe capabilities. However if using Azure managed service is not a strict requirement you can use RabbitMQ on a VM (or a cluster for high availability) as well.
UPDATE: Your response means you want a managed service. There are 2 options - if you want to go with RabbitMQ but do not want to manage the infrastructure you can go for 3rd party service provider like CloudAMQP who will manage it on your behalf. The other option is to go for Cloud native messaging - meaning if you are on Azure you change your messaging service to Azure Service Bus. This would mean changing you code as well.
Right now, on IoT Hub there is an information that limit for messages per day 8000. I would like to ask you about any patterns which are being used in Azure.
I am curious if I am able to hit to Azure with some service outside Messages in order to prevent it from being overloaded by big amount of data, or save some confidentiality for this service.
For example, I would like to store some data from given service to Messages that are not being confidential and other data by using some WebSocket or any Rest protocol. I think that there are some patterns that serve that scenarios.
Does anyone has experience with that kind of situation?
Not everything needs to go through IoT Hub. IoT Hub is great for two way communication to/from IoT devices. You could also look at Event Hubs for ingestion from devices that don't need two way comms. We have a write up on the differences here Connecting IoT Devices to Azure: IoT Hub and Event Hubs.
I have an Azure IoTHub with thousands of devices registered. These devices communicate through a Telco provider who sends messages through an Azure Storage Queue. This Storage Queue triggers an Azure Function which needs to parse the messages and Send an Event to the IoTHub as below.
Currently, we use the Azure IoTHub SDK to create a DeviceClient for each payload and we send the event. Because the DeviceClient represents a device in the IoTHub and is carrying the context of the source of the events, we are having to recreate a device client for each event. This quickly exceeds the threshold of the number of Connections allowed on Azure Functions.
We have tried using the IoTHub Output bindings for Azure Functions, but could not get to work and I do not think it would work because we need to make sure that the events get to the IoTHub with the right context (messages are sent by the right device).
What's the right way to solve this? Can the connections to the IoTHub be reused? Should we abandon Azure Function in favour of something else?
I assume that Telco is some kind of custom device management solution(vendor lock solution), that can also communicate with the device and receive the device telemetry, and eventually forward it to the specified endpoint, correct?
If I may ask and if my assumption is correct, why do you need to deliver the events to IoT Hub, if you are not managing Telco devices through IoT Hub(the arrows on your diagram are only in one direction)?
Using the IoT Hub just as a message broker for essentially cloud-to-cloud communication is not beneficial if that is the only purpose. Also conceptually what you described is cloud-to-cloud communication, and IoT Hub is intended to be used for devices.
Here is what I would do. Setup the API Management(or http triggered Azure Function) as a front door for Telco and pass the messages to the Event Hub.
You can choose here to pass request body for example where your telemetry data is - I assume again.
Keep the IoT Hub, and setup the routing to previously created Event Hub.
Now, in case you have devices that are not vendor locked and that can talk directly to IoT Hub, messages will be re-routed to Event Hub. Also Telco device messages will be routed to exactly the same Event Hub.
Now you can have for example Azure Stream Analytics that can analyze data stream just from the Event Hub, and for both, Telco devices and potentially non-Telco devices.
After trying a few things, I ended up moving away from using the SDK for pushing messages to IoT Hub. This is because the SDK uses AMQP, and creating a DeviceClient for each payload is not viable.
We switched to using HTTPS instead to push the messages to IoT Hub and using HttpClientFactory, we are able to do connection pooling.
I thought I would put this here in case someone has the same issue.
Here is an example of the Http request to send message to IoT Hub
Host: https://<iothubname>.azure-devices.net/devices/<deviceId>/messages/events?api-version=2018-06-30
Authorization: SharedAccessSignature sr=<iothubname>.azure-devices.net&sig=abc123;12344iweoippweruea=iothubowner&se=1570574220
Body: <normal Interval or alarms payloads> // example {"deviceid": "abc", "hello": "world"}
Lastly, thanks #kgalic for the answer but your suggestion would not work. This is not pure B2B integration. Our implementation have to allow for both devices connecting directly to the IoT Hub and devices connecting through the Telco. This is why every device needs to have its own identity and digital twin.
So, I am working on an IoT solution on Azure, we have been using a partner solution where we had the partner's devices linked to his cloud solution that exposes the data to us Via REST services. Right now we want to have our own IoT Cloud Solution on Azure.
At first, I am planning to build a Bridge between our IoT Solution and the partner's cloud solution via its REST Services that will link to our IoT Hub in order to ingest the data to our cloud.
Also, the data will not be only telemetry data but we'll have to send commands as well to those devices.
My question: I would like to know what would be the appropriate technology/solution to use a gateway (Data Grid, Azure Function, Azure WebJob)
The numbers in the picture represent the step that I am considering to tackle this problem.
1- First we are implementing an Application gateway that will have to get the data from the partner's system and sending commands to their system. It will allow us to first build the other components of our system and make sure that it can handle what is in place right now.
2- Second, the partner's devices will connect directly to a device gateway that is connected to our IoT Hub. In this case, we will not be using the gateway made in 1 anymore.
3- Finally, we will have our own devices connected to our IoT Hub, the partner's devices will always be connected to our IoT Hub via the gateway built in 2.
Let me try to answer your questions in the order you have asked.
For application gateway, where you are trying to pull data through
REST, you can use Azure functions and then you use Cosmos DB or any
storage to save data. I see , after getting device data from Partner
network, you are routing it to IoT-Hub (I would not say, its
incorrect), however once we pull data through Rest, we can directly
put into DB. So my Answer is to use Azure functions to pull data
from Partner solutions and put into DB.
If partner device is capable of running Azure IoT sdks or can be
provisioned to send data to IoT Hub directly, this will ease lot of
things and you would be able to send D2C and C2D messages easily.
further, here you can route data to DB by using configuration from
IoT Hub.
For your devices you can use IoT Hub Directly or can use Azure
IoT Edge (device gateway as you pointed ), both are fine , depends
on use case and also if we want to do some edge computation or
analytics at device side. And one important suggestions, use Azure
functions where ever you find that you have to integrate devices
data through Rest. Most cost effective in such scenarios.
Let me know if it clears your doubts.
After some time working on the subject, I did implement an AZURE Function app for the following reasons :
Supports Continuous Deployment and Integration Even though Azure Functions is serverless architecture, it still supports Continuous Deployment and Continuous Integration
Capabilities for implementing code - Being event-driven, the application platform has capabilities to implement code triggered by events occurring in any third-party service or on-premise system.
Compute-on-demand: This delivery model ensures that computing resources are available to the users as per their demand.
I have also used Azure Table Storage as database storage technology.
I'm simply trying to work out how best to retrieve messages as quickly as possible from an Azure Service Bus Queue.
I was shocked that there wasn't some way to properly subscribe to the queue for notifications and that I'm going to have to poll. (unless I'm wrong in which case the documentation is terrible).
I got long polling working, but checking a single message every 60 seconds looks like it'll cost around £900 per month (again, unless I've misunderstood that). And if I add a redundant/second service to poll it'll double.
So I'm wondering what the best/most cost efficient way of doing it is.
Essentially I just want to take a message from the queue, perform an API lookup on some internally held data (perhaps using hybrid services?) and then perhaps post a message back to a different queue with some additional information .
I looked at worker roles(?) -- is that something that could do it?
I should mention that I've been looking at doing this with node.js.
Check out these videos from Scott Hanselman and Mark Simms on Azure Queues.
It's C# but you get the idea.
https://channel9.msdn.com/Search?term=azure%20queues%20simms#ch9Search
Touches on:
Storage Queues vs. Service Bus Queues
Grabbing messages in bulk vs. one by one (chunky vs. chatty)
Dealing with poison messages (bad actors)
Misc implementation details
Much more stuff i can't remember now
As for your compute, you can either do a VM, a Worker Role (Cloud Services), App Service Webjobs, or Azure Functions.
The Webjobs SDK and Azure Functions bot have a way to subscribe to Queue events (notify on message).
(Listed from IaaS to PaaS to FaaS - Azure Functions - if such a thing exists).
Azure Functions already has sample code provided as templates to do all that with Node. Just make a new Function and follow the wizard.
If you need to touch data on-prem you either need to look at integrating with a VNET that has site-to-site connectivity back to your prem, or Hybrid Connections (App Service only!). Azure Functions can't do that yet, but every other compute is a go.
https://azure.microsoft.com/en-us/documentation/articles/web-sites-hybrid-connection-get-started/
(That tutorial is Windows only but you can pull data from any OS. The Hybrid Connection Manager has to live on a Windows box, but then it acts as a reverse proxy to any host on your network).
To deal with Azure ServiceBus Queue easily, the best option seems to be Azure Webjob.
There is a ServiceBusTrigger that allows you to get messages from an Azure ServiceBus queue.
For node.js integration, you should have a look at Azure Function. It is built on top of the webjob SDK and have node.js integration :
Azure Functions NodeJS developer reference
Azure Functions Service Bus triggers and bindings for queues and topics
In the second article, there is an example on how get messages from a queue using Azure Function and nodejs :
module.exports = function(context, myQueueItem) {
context.log('Node.js ServiceBus queue trigger function processed message', myQueueItem);
context.done();
};