From an architectural point of view, an IoT platform presents significant challenges:

• The amount of data acquired, ingested, and processed requires high bandwidth, storage, and computational capabilities
• The devices are distributed across a vast geographical area
• Businesses require their architecture to be evolutionary so that new services and capabilities can be added daily to deploy to customers

More so than other standard applications, flexibility and scalability are vital to a typical IoT platform. However, IoT and I-IoT both have an interesting benefit, which is the fact that the relationships between the actors are weak. In the IoT platform, signals are independent. Data sharding can be used for storing purposes and parallel computation can be used to improve computational performance. On a typical ERP or e-commerce application, the data should remain centralized and connected. The following diagram highlights these differences:

This feature of the IoT platform makes it more similar to a NoSQL database than a SQL database, a microservice-based application than a monolithic application, and a cloud- or fog-based system than a centralized system. Therefore, IoT is more related to an ecosystem of services and applications than to an actual application.