A High Level Look at the World Form Reference Model – To start, we need an introduction to the seven layers of this World Forum Reference Model. The introductory walk through is worthwhile as there are some subtleties that are worth noting. However, I will keep this necessary, term setting, introduction brief, as we will go into more detail on each section, and most readers will be familiar with the conceptual layers of the model.
Starting from the edge:
Physical Devices and Controllers – The model calls this layer the “things” of the internet of things. Unfortunately, this layer is a little ambiguous. One one hand, the “things” are the assets being managed. From a system design perspective, the “things” are the sensors and devices that are directly managed by the IoT architecture. These two are not often identical, at least not yet, as many assets (“things”) under management will not yet have integrated sensors and Edge Node intelligent elements integrated. So, we should think of this layer as consisting of the “things” themselves and the sensors and Edge Node devices connected to them, as well as a more modern class of “things” with integrated sensors and Edge Node functionality.
It is worth noting that in the later case, and where connecting to existing field assets, there can be significant design effort to connect the sensors and Edge Node intelligent hardware as well as in mapping these systems to any management or intelligence that may exist in legacy assets (unless they are just “dumb” assets that need instrumentation).
An important IoT concept, Edge Intelligence, to allow low latency reaction to field events and to allow higher levels of autonomy and distributed processing, needs to be implemented at this layer.
Connectivity – This layer spans from the the “middle” of an Edge Node device up through transport to the cloud. Many alternatives can be used for communications and this layer includes the mapping of field data to the logical and physical technologies used as well as the backhaul to the on premise or cloud and the next layer, Edge Computing.
In deployment, this layer can use a single solution or multiple technologies, depending on the need. Field Area Networks (FANs) alternatives can include wired, cellular, LPWAN, and many other wireless options, as well as multi-tired solutions, and can be build out of private, public, or a mix of private and public transport solutions.
Edge Computing – The next layer in the World Forum Model architecture is Edge Computing, or more properly “Cloud Edge” or “Cloud Gateway” computing. Required to some degree in any IoT system this layer interfaces the data and control plains to the higher layers of cloud, SaaS, or enterprise software layers. Protocol conversion, routing to higher layer software functions and even “fast path” logic for low latency decision making will be implemented at this layer.
Data Accumulation – Given the Velocity, Volume and Variety that IoT systems can provide it is essential to provide incoming data storage for subsequent processing, normalization, integration, and preparation for upstream applications. While part of the overall “data lake” architecture, this layer of the architecture serves the intermediate storage of incoming storage and outgoing traffic queued for delivery to lower layers. This layer may be implemented in simple SQL or may require more sophisticated Hadoop & Hadoop File System, Mongo, Cassandra, Spark or other NoSQL solutions.
Data Abstraction – In the data abstraction layer we “make sense” of the data, collecting “like” information from multiple IoT sensors or measurements, expedite high priority traffic or alarms, and organize incoming data from the data lake into appropriate schema and flows for upstream processing. Similarly application data destined for downstream layers is reformatted appropriately for device interaction and queued for processing.
A key architecture element for larger high performance deployments is a publish / subscribe or data distribution service (DDS) software framework to simplify data movement between Edge Computing, Data Accumulation, Application Layer, and User Processes. Whether this is a high performance service or a simpler message bus this infrastructure simplifies implementation and improves performance for all but the simplest applications.
Application Layer – This layer is self explanatory and is where control plane and data plane application logic is executed. Monitoring, process optimization, alarm management, statistical analysis, control logic, logistics, consumer patterns, are just a few examples of IoT applications.
Collaboration and Processes – At this layer, application processing is presented to users, and data processed at lower layers is integrated in to business applications. This layer is about human interaction with all of the layers of the IoT system and where economic value is delivered. The challenge at this layer is to effectively leverage the value of IoT and the layers of infrastructure and services below and leverage this into economic growth, business optimization and/or social good.