From the perspective of the I-IoT data flow, DCSes are very similar to PLCs, so all the previously analyzed strengths and weaknesses also apply in this case. There are just a few differences to underline, which are listed as follows:

• The architecture of the DCS is natively organized hierarchically. This means that in industrial plants where the automation systems mainly use DCSes, we don't find many of them and they are typically supplied by the same vendor. For example, in a medium-sized refinery, we might have three or four DCSes that manage the whole plant.
• The DCS plays the double role of controller and SCADA system. Many also have an integrated Historian natively connected to the DCS controllers. This helps to select the data to be gathered since they are usually organized and structured according to a data model. This means we have to deal with less variety and complexity compared to using PLCs.
• The DCS, like the PLC, manages and collects thousands of sensors, actuators, internal calculations, and derived values and statuses. Typically, its scan rate is between 100 and 1,000 milliseconds. This is lower than the scan rate of a PLC but in most cases, the data is not connected by linking to its native protocol, but instead through its OPC server. Any OPC server is unlikely to provide data to its OPC clients with a refresh time under 1,000 milliseconds. This means that from the perspective of the I-IoT data flow, it doesn't really matter whether the underlying device is a DCS or a PLC when gathering data from an OPC server.