CPS experimentation suffers from several limiting factors, including the following:
- Simulation tools are often developed in isolation, using a wide variety of architectures and platforms. A simulation tool developed in one environment can't be easily used in another environment without custom-developed adapters.
- Some CPS experiments are done in operational systems (e.g., on a live electric grid or in a moving test vehicle). Because these operational systems may be providing critical functions in real time, these experiments are limited by the severe constraints required to ensure that experiments and testing do not affect “trustworthiness.” (“Trustworthiness” is a concept that encompasses security, safety, reliability, resilience, and privacy.)
- Some CPS experiments include proprietary components that owners don't want to share or reveal.
- Some CPS experiments involve equipment that is unique or can’t be collocated.
Because of these types of constraints, it is difficult to study the general applicability of CPS concepts and technologies intended for implementation across multiple domains and in varied applications.
As scientists and engineers need to conduct sophisticated experiments involving systems and "systems of systems" of ever-increasing complexity, it will be very useful to have a multi-domain CPS testbed with remote federation functionality. Such a testbed should be:
- reproducible; and
- and scalable.
UCEF has been developed to address these concerns.