Losant Edge Compute is a suite of functionality within the Losant platform that allows for deploying workflows to your connected devices and executing those workflows on the device itself. This is in contrast to application workflows, which execute in Losant’s platform.
Fully utilizing Losant Edge Compute brings a number of benefits to your application.
The ability to make decisions on the device – and not in the cloud, which depends on an internet connection (much less a fast one) – can be critical to an IoT application. For example, if you have a vibration sensor attached to a machine, and that machine should shut down if vibration levels spike (a clear indication of a fault), even a one-second difference can mean the difference between a simple repair and a new piece of equipment.
This difference is achieved by acting on the value of that sensor and making the “shut-down” decision locally on the device via an edge workflow. The alternative is sending the sensor readings through the internet up to the cloud, where the evaluation of the data takes place in an application workflow. Then, a command to shut the machine down must be sent back through the internet down to the connected device.
If your devices are deployed in environments where internet service is intermittent, you can use Losant Edge Compute to accumulate state reports on the device. Then, the next time the device connects, those reports can be sent up to the cloud platform.
Your devices and sensors can potentially generate hundreds of data points per second, and it’s likely you only care about long-term averages and/or anomalies in these data points. By filtering the sensor values at the local level, you can dramatically reduce data noise within your application, cut the number of payloads generated and save battery life by reducing bandwidth.
Much like with the Workflow Engine itself, Edge Compute allows less-technical end users to make behavior changes within their connected devices without having to work in a coding environment. Then, updating, versioning and deploying the changes is all done through Losant’s web interface, and since all changes require creating a version, any mistakes can quickly be rolled back by deploying an older, stable version.
To get started with Edge Compute, you will have to follow these steps. More detailed information is available within the respective documentation for each section.
- In your application, create a new device and set its device type to “Edge Compute”. Only devices of this type can receive workflows to run locally from the Losant platform.
- Create an access key and secret for your new device. While we always recommend generating a new key/secret pair per device, this step may not be necessary if you already have a key and secret configured to work for all devices in your application, or to work for all devices with a tag that you applied to your device during creation.
- Deploy the Edge Agent to your new device and configure it using the device ID, key and secret from your previous steps. The Edge Agent software package is necessary to receive edge workflows from Losant and to run them on the device.
- Create a new workflow of the “Edge Workflow” type. Make sure the Target Agent Version is less than or equal to the version of the Edge Agent you deployed to the device in the previous step.
- Set up your workflow using any of the nodes in the palette and save your work.
- When it is ready, deploy a version of your workflow to your Edge Compute device(s). If the device is connected to Losant’s MQTT broker – or when it connects next – the edge workflow version will deploy to the device and it will begin executing.
Installing or upgrading the Edge Agent requires physical access to the device. After that, management of all edge compute functionality - device configuration, workflow setup and deployment schedules – takes place within the Losant platform and updates are pushed down to the devices over the internet.