State machines for large scale computer software and systems (1512.06788v4)

Abstract: This paper introduces techniques for specifying behavior, architecture, and abstract properties of large scale computer software and hardware purely in terms of ordinary deterministic state machines. The goal is to be able to work with specifications ranging from operating systems to databases and real-time control. State machines with output are represented as maps from finite sequences of events to outputs so that f(s) is the output in the state reached by following s from a initial state. Composite sequence maps correspond to state machine products. The methods used here can specify or constrain both system behavior and system design. Motivating examples presented range from simple counters to distributed consensus algorithms and real-time circuits. The approach is intended to facilitate "back of the envelope" descriptions of devices and software and also to allow for detailed hierarchical specifications of behavior and architecture. The mathematical approach is based on both primitive recursion on sequences \cite{PeterComputer} and automata products with feedback \cite{Gecseg}, adapted to Moore type state machines. No formal methods or other metamathematical techniques are employed and although parallel and concurrent composite systems are easy to specify, it is not necessary to make any particular communication scheme primitive. State machines are not augmented or extended - even the most complex composite systems are modeled by ordinary deterministic state machines which have a rich mathematical theory related to semigroups.