Originally published on LinkedIn: https://www.linkedin.com/pulse/iot-requires-modeling-reality-ron-lake 
While a great deal of the focus of the Internet of Things (IoT) is on devices and the assignment of IP addresses to objects, it is increasingly recognized that modeling the reality in which these devices are embedded is a critical component of most IoT applications. We have referred to this in previous blogs and presentations as the ‘device context’ and it is embraced in Cisco’s phrase the ‘Internet of Everything’.
Since we can anticipate that multiple applications may use the same sensors or actuators, multiple different models of reality may overlap and intertwine. For example, suppose we have components that are installed in the field by heavy equipment, such as road mats in oil and gas exploration. Our initial application may incorporate sensors in the mat that measure the mat’s location and degree of wear, as well as sensors in the vehicle installing the mat that measure its location, speed, etc. This application focuses on recording where mats are installed and the heavy machine time required per mat installed. These sensors, perhaps combined with in-vehicle sensors such as seat pressure and operator ID, may also help to track the hours the machine is in use by a particular operator as part of a second application that is focused on machine usage and operator performance. In the first case our model of reality focuses on the mat and the heavy machines used to install them, while in the second it focuses on the heavy machines and their operators.
With such multiple applications it will likely make sense to maintain models in a manner that enables the models to be interconnected with one another. This could be achieved through the use of common objects types and global identifiers (e.g. road segment type, road segment identifier), and through the relationships between these real world object types and one another (e.g. vehicle drivesOn road_segment), and between these real world objects and the IoT devices (speed_sensor installedOn vehicle).
Over time, we can anticipate the development of real world models for significant object collections such as areas in cities, oil and gas fields, transportation systems and networks, and vehicles and machines. The beginnings of such models already exist today for many domains, such as CityGML for urban areas, PPDM for oil and gas wells, AIXM for aviation and airports, and PODS for oil and gas pipelines, as well as the more specific models developed by corporations or government in particular domains. We can look forward to a time when an IoT device can first be “installed” in our model of reality prior to being installed in the real world. This will require interoperability of the data models on which these models of reality are based, which coincides with requirements for such models in Virtual and Augmented Reality.
Many will complain that the creation and maintenance of such models is an undoable task, that the volume of data is too large, and the maintenance of that data is too difficult. However, we don’t have to create such models all at once, nor do they have to live in only one place or on one machine, nor must they encompass all the things in the world from the get go. Registries of Things will emerge in particular domains (e.g. pipeline monitoring), and rapidly provide support for activities like pipeline construction and vehicle utilization.
Creating these models of reality will be a challenging task, but I believe it is a task that the OASIS/OGC Registry Platform (CSW-ebRIM) is up to. So let’s get started!