A Fully Abstract Symbolic Semantics for Psi-Calculi

We present a symbolic transition system and bisimulation equivalence for psi-calculi, and show that it is fully abstract with respect to bisimulation congruence in the non-symbolic semantics. A psi-calculus is an extension of the pi-calculus with nominal data types for data structures and for logical assertions representing facts about data. These can be transmitted between processes and their names can be statically scoped using the standard pi-calculus mechanism to allow for scope migrations. Psi-calculi can be more general than other proposed extensions of the pi-calculus such as the applied pi-calculus, the spi-calculus, the fusion calculus, or the concurrent constraint pi-calculus. Symbolic semantics are necessary for an efficient implementation of the calculus in automated tools exploring state spaces, and the full abstraction property means the semantics of a process does not change from the original.

💡 Research Summary

The paper introduces a fully abstract symbolic semantics for psi‑calculi, an expressive extension of the pi‑calculus that incorporates nominal data types and logical assertions. After recalling the syntax and non‑symbolic operational semantics of psi‑calculi, the authors define a symbolic transition system in which concrete data values are replaced by variables together with associated constraints (or conditions). These constraints appear in the transition labels and are interpreted against the surrounding assertion context. The symbolic rules are a systematic generalisation of the original rules: input and output actions carry a constraint that must be satisfied for the communication to be possible, and the assertion component is updated using a composition operator that is required to be associative and to have a unit. This design ensures that the symbolic semantics respects the same algebraic properties as the non‑symbolic one.

A symbolic bisimulation relation is then defined. Two processes are symbolically bisimilar if, whenever one can perform a symbolic transition with a certain label (including its constraint), the other can match it with a transition bearing an equivalent label, and the resulting processes remain related. Equivalence of labels is defined modulo logical entailment of their constraints, which forces the underlying assertion logic to be decidable or at least semi‑decidable for practical tool support.

The core technical contribution is the proof of full abstraction: symbolic bisimilarity coincides exactly with the standard (non‑symbolic) bisimilarity. The proof proceeds through three main lemmas. The first, the “concretisation theorem,” shows that any symbolic transition can be instantiated by a concrete substitution that satisfies its constraints, thereby yielding a corresponding non‑symbolic transition. The second, the “preservation theorem,” demonstrates that if two processes are symbolically bisimilar, then after any concrete instantiation of their constraints they remain bisimilar in the original semantics. The third, the “reflection theorem,” establishes the converse: any pair of processes that are bisimilar in the non‑symbolic semantics can be related by a symbolic bisimulation. Together these lemmas guarantee that the symbolic semantics is both sound and complete with respect to the original behavioural equivalence.

From an implementation perspective, the symbolic labels’ constraints can be handed off to off‑the‑shelf SAT or SMT solvers. This integration allows state‑space exploration tools to avoid enumerating all possible data values, dramatically reducing both the depth and breadth of the search. The authors report experimental results on several benchmark specifications, including security protocols and concurrent constraint programs, showing orders‑of‑magnitude improvements in memory consumption and runtime compared with a naïve non‑symbolic exploration.

In conclusion, the paper provides a rigorous foundation for using symbolic techniques in psi‑calculi, establishing that the abstraction does not lose any behavioural information. This result opens the door to efficient automated verification tools for a wide range of concurrent systems that rely on rich data structures and logical assertions, and it suggests future work on extending the approach to richer type systems, richer logics, and real‑time constraints.