Virtual Sensors Deployment
The only thing you need to worry during deployment, is selecting which Virtual Sensor to deploy. It is up to the Platform to analyse your request, and deploy it on your physical infrastructure. To accomplish this task, the Platform takes under consideration multiple metrics such as performance requirements and current status of the infrastructure to generate an optimal distribution of the workload, considering all Virtual Sensors you have selected.
To create a new deployment select NEW DEPLOYMENT. From the menu select all Virtual Sensors you want to deploy, and click on DEPLOY.
This is all the effort required to you to create a new deployment.
To generate a new deployment you can just create a new deployment with zero Virtual Sensors. In this way the Platform will set all the nodes to a base configuration.
The deployment process goes through multiple steps, starting from compiling.
After applying the Workload Distribution logic, the Platform moves to generate, for each portion of the algorithm, a representation that can be deployed on the target node. Hence, during the compiling step, for each code portion that will be delivered to a sensor node, the platform automatically compiles a specific firmware; while for each code portion that will be delivered to a local or cloud node, it generated a container that can be executed on the specific node.
After each code portion is compiled the platform moves to the real deployment and performrs Over The Air (OTA) firmware updates for the sensor nodes, and updates for the local and cloud nodes, and wait for all these updates to complete.
During these steps, you have always the possibility to check the status of the deployment by clicking RUN CONNECTIVITY CHECK and using the drop-down menu of the deployment to check the status of each node.
Once all the steps have been completed and each node updated, the deployment status becomes ready, and meaningful data is published to your cloud resource specified during project setup.
The Platform offers the possibility to execute deployments dynamically, based on specific conditions that can happens. The entrypoint for the creation of a dynamic deployment is a Deployment Hook.
The creation of a deployment hook works as the creation of a deployment, and generates in output a link to the Platform that when accessed triggers the specified deployment.
An example usage of the deployment hook are the rules inside Azure Central. With rules you can triggers operations based on the outputs published by the Virtual Sensor, so that the deployment hook is triggered.
Once a deployment hook is triggered, you can see in the Platform a new deployment flagged as automatic