Towards the distributed edge-An IoT review

The internet of Things (IoT) architecture was originally envisaged as a two-layer technical platform, with sensors collecting data at the edge with minimal compute requirements, solely to prepare and transporting the data to a centralized or cloud based infrastructure for processing. This model is suitable in some scenarios, for example where data is being stored for historical, regulatory or trending usage however in other use cases such as health monitoring for acute illness or autonomous vehicle computer vision the latency in transporting this data to a remote location for processing may cause latencies that would seriously affect performance of the application. There are many different and sometimes overlapping definitions of IoT topologies being discussed within industry. This paper reviews this original topology of an IoT solution, and different techniques and layers available to alleviate the issues inherent of the original paradigm, and how a new method of defining at these topologies is gaining speed.