Using the Actor Model and Akka.NET for IoT Systems

In additional to cloud-based offerings such as Microsoft’s Azure IoT Suite, it’s also possible to create representations of Internet of Things devices using the Actor Model and Akka.NET. For example, in addition to hosting an Akka.NET actor system in the cloud, you can also have them running on-premises; for example you may have an actor system monitoring an underground railway network for a city. In this case you may already have the infrastructure in place and decide that hosting in the cloud is unnecessary or risky, and deploy to servers geographically close to the underground with some disaster recovery/backup servers located elsewhere.

The Actor Model is a good fit for IoT scenarios due to the inherent concurrency controls, fault-tolerance, and performance/scalability. Akka.NET is a “toolkit and runtime for building highly concurrent, distributed, and fault tolerant event-driven applications on .NET & Mono”[1].

In the Actor Model, the smallest unit of computation is the actor. An actor can receive messages from other actors, perform computations, manage their state, and send messages to other actors.

For IoT scenarios, actors can model the system in a number of ways. For example, for every physical device in the real world there can be an actor instance to represent that device. This means you may have many actor instances, one for each physical device. When a sensor registers a new reading (temperature, proximity, pressure, etc.) this sensor can communicate with the actor responsible for it, the actor receives a message and can update its internal state to represent the latest reading. The actor representing the device can also respond to another type of message to access the last updated value. The device actor may also be responsible for sending messages to the physical device to update it’s configuration (for example) or this functionality may be broken down into its own actor definition depending on the exact requirements. The device actor may communicate with another actor that is purely responsible for handling the network interfacing required to communicate with devices. In this way if the “network actor” crashed and restarts the actors representing the device don’t crash and lose their state.

Groups of “device actors” can created with the supervision hierarchy that the Actor Model provides to protect groups of actors/devices from crashing other groups of actors/devices. These hierarchies can also provide device management semantics such as the registering of a new device that has just come online. Actors can also represent “abstract” concepts such as the creation of a dedicated actor to perform querying of groups of actors.

To learn more about using Akka.NET with IoT scenarios, check out my Representing IoT Systems with the Actor Model and Akka.NET Pluralsight course.

You can start watching with a Pluralsight free trial.


Add comment