The industrial automation industry offers several classes of edge computing reliability. For nominal reliability, simplex PLCs provide a very good solution. Distributed Control Systems (DCS) offer much high reliability in either a simplex or redundant configuration. Safety critical PLCs offer even higher reliability through the use of a triple modular redundant design (TMR).

However, once a design has been implemented it becomes very expensive to change it – and if actual failure rates do not measure up to the manufacturers published figures, the plant could be exposed to unnecessary operational risk. Furthermore, flexibility to choose different DCS configurations for different parts of a plant is limited to only high reliability (simplex) and ultra-high reliability (redundant) – which may force the designer to over-specify reliability in sections of the plant with lower operational risk. Finally, multiple edge computer equipment classes (i.e. Simplex-PLC, DCS and TMR-PLC) impose a maintenance and cost burden on an operating facility for additional training, spare parts and external commercial interfaces. They also create interoperability challenges that may lead to stranded data and a partitioned operational view of the plant.

This presentation will offer a unified open solution that allows a user to specify the reliability target of any edge computing cluster through configuration alone. A proposed methodology based on open industry standards will also be shared to manage low-latency bumpless failover to any backup computer located anywhere in the system. The failover is managed entirely through communications middleware which redirects data flows in accordance with the user’s desired quality-of-service.