Synthesizing Stream Control

Felix Klein

For the management of reactive systems, controllers must coordinate time, data streams, and data transformations, all joint by the high level perspective of their control flow. This control flow is required to drive the system correctly and continuously, which turns the development into a challenge. The process is error-prone, time consuming, unintuitive, and costly. An attractive alternative is to synthesize the system instead, where the developer only needs to specify the desired behavior. The synthesis engine then automatically takes care of all the technical details. However, while current algorithms for the synthesis of reactive systems are well-suited to handle control, they fail on complex data transformations due to the complexity of the comparably large data space. Thus, to overcome the challenge of explicitly handling the data we must separate data and control.

We introduce Temporal Stream Logic (TSL), a logic which exclusively argues about the control of the controller, while treating data and functional transformations as interchangeable black-boxes. In TSL it is possible to specify control flow properties independently of the complexity of the handled data. Furthermore, with TSL at hand a synthesis engine can check for realizability, even without a concrete implementation of the data transformations. We present a modular development framework that first uses synthesis to identify the high level control flow of a program. If successful, the created control flow then is extended with concrete data transformations in order to be compiled into a final executable.

Our results also show that the current synthesis approaches cannot replace existing manual development work flows immediately. During the development of a reactive system, the developer still may use incomplete or faulty specifications at first, that need the be refined after a subsequent inspection. In the worst case, constraints are contradictory or miss important assumptions, which leads to unrealizable specifications. In both scenarios, the developer needs additional feedback from the synthesis engine to debug errors for finally improving the system specification. To this end, we explore two further possible improvements. On the one hand, we consider output sensitive synthesis metrics, which allow to synthesize simple and well structured solutions that help the developer to understand and verify the underlying behavior quickly. On the other hand, we consider the extension of delay, whose requirement is a frequent reason for unrealizability. With both methods at hand, we resolve the aforementioned problems and therefore help the developer in the development phase with the effective creation of a safe and correct reactive system.

PhD Thesis.