Issue link: https://www.e-digitaleditions.com/i/1347119
April/May/June 2021 I 9 Consider what this looks like for reading analog data through an IO card and sending it over an Ethernet link to another device (Figure 2). The analog sensor constantly provides a signal that reflects its current condition, such as temperature. On a set interval, the IO card reads the signal and converts it to usable units. Once this is done, it must decide on the application layer framework in which to deliver the data and its context. There are many options. It could use an internet of things (IoT) protocol such as MQTT (originally an acronym for MQ telemetry transport) or another open ISO standard pro- tocol supported by existing industrial vendors. From there, the IO card decides where the data needs to be sent, puts the data into a packet, and delivers it over the transportation layer. Much of the above process must be programmed into the IO card either from the factory or by engineers at system commissioning. The process is repeated at each interval of device reading. While this architecture is described for analog sensors, much the same can be said of other systems that use legacy serial protocols, as well as emerging wireless IoT sensor systems that are sold as kits. In the case of wireless IoT, the kitting approach is required for sensors to communicate their data to servers. A gateway is required to interpret the data communi- cated over a wireless protocol and to forward as a TCP/IP packet through the traditional network. These gateways are typically custom designed to work with proprietary wireless communications or to reach the objective of reducing engineering implementation effort as much as possible. Entering the age of digitalization, much of the data that is being automatically connected can- not be used straightaway. Currently, there is a lot of effort to take the data gathered at edge devices all over the world and convert them for use. It stands to reason that efficiency can be gained if data is created and communicated from the source in the same form in which it will be consumed at the ERP, MES, or other servers. Consider the above scenario but now with a smart architecture in place that communicates by Ethernet all the way down to the edge temperature sensor or actuator. In the prior example, the data was collected by an analog sensor but was never made acces- sible outside of the cell. In this current example, the sensor is on an IP network and can be accessed through its IP address rather than requiring data accessibility that requires the PLC to provide a tag. The sensor can publish the data to any location where it is needed through FIGURE 2: How analog and fieldbus sensors send data in context of the OSI model. Analog Sensor 7 Sensor Calibration 6 5 Analog Temp Reading 4 3 2 0-10V Signal 1 IO Card Read & contextualize Convert to (e.g., MQTT) Convert to temperature Where are you sending the data? Receive 0-10V Signal Packetize and Send Controller Interpret & use data Receive