In manufacturing, oil and gas, and transportation, OT stacks have been assembled, generally, from component suppliers: programmable logic controller (PLC) vendors (for example Siemens, Rockwell Automation, Schneider Electric), supervisory control and data acquisition (SCADA) providers, network equipment vendors (e.g., Cisco), and separate analytics (e.g., OSIsoft before its acquisition by AVEVA). McKinsey’s surveys in these sectors notes that fragmented stacks are among the barriers to scaling IoT, particularly in asset-heavy sectors like chemicals and energy.
Each element (OT, comms, attenuation & control, data, analysis and presentation) introduces a degree of licensing cost and/or integration work. Gartner’s IoT Hype Cycle and Magic Quadrant commentary for 2019-2023 singles out integration complexity as one reason why scaling projects often see costs rise in line with growth of scope – that is, economies of scale are at least partially negated by integration overheads..
At a practical level, multi-vendor solutions can run into issues around data models differing between systems, stakeholders each requiring custom data analysis and associated interface, and upgrades to one element in the stack can be predicated on another. Plus, there’s the issue of security management becoming more complex. The US Department of Energy and ENISA have both said that heterogeneous vendor environments increase the attack surface and complicate patch management in many types of industrial control systems, for example.
A tertiary source of cost can be found in the process of managing multiple contracts. This consumes procurement and legal resources, and represents a silent cost not necessarily attributed to factory floor operations.
More recently, connectivity and device management standards have simplified interoperability, like MQTT, OPC UA, and open APIs. The Eclipse Foundation’s surveys of IoT developers show MQTT and OPC UA among the most adopted protocols, particularly in manufacturing and energy. Combining interoperable hardware and software (like edge processing from Litmus Automation, ClearBlade, or Crosser) with cloud services for aggregation and analysis is easier as – at least in theory – local integrations use the same tooling as is common in cloud-native environments. APIs allow integration with, for example, enterprise systems (SAP, Oracle ERP, etc.), for visualisation and reporting, regardless of where they may be hosted.
