IoT Edge Architecture


Last days, a colleague representing Galeo’s client ask me about the rationale of the IoT Edge architecture designed a couple of years ago. Even if there are many versions of IoT reference architectures, it’s always good to remember the principles behind we design our solutions.

Principles of architecture design.

Concentration strategy. Field IoT Edge gateways are proposed to deploy when a concentration strategy is chosen. IoT Edge Gateway aggregates information and connectivity capabilities to integrate lower-level equipment downstream. A concentration strategy is almost always recommended for handling the following situations:

  • Translation of field protocols. At the field level, there are several systems, from different manufacturers with very different communication protocols and data structures. An IoT Edge Gateway performs the functions of “dialoguing” with these systems through various languages (REST API, MQTT, Modbus RTU, Modbus TCP, Proprietary protocols, etc.).
  • Isolation of OT networks from IT networks. An IoT Edge Gateway, typically, acts as an isolation element for OT networks. It is considered good practice to decouple OT networks from IT networks for performance, skills, security, etc.
  • SW and IoTOps. With an IoT Edge Gateway strategy, the life cycle of the SW (deployed on the field) is managed in a much more scalable and maintainable way. Besides, there is a computing-ready platform close to the control and action elements where to deploy ad-hoc functional SW.
  • Local processing and storage capabilities. The IoT Edge Gateway enables the deployment of SW for data processing and local (process) functionalities. For example, an SW module harmonizes and enriches the messages coming from different vendor systems. Local storage can also be deployed to persist device data in the field. This also serves as a backup measure.
  • Flexibility for new use cases. The IoT Edge Gateway provides flexibility to adapt the solution to new future use cases. For example, if computer vision strategies are implemented for image analysis of security cameras it is necessary to have an element such as the IoT Edge for the local management of these datasets.