Synthesis of Component and Connector Models from Crosscutting Structural Views

We present component and connector (C&C) views, which specify structural properties of component and connector models in an expressive and intuitive way. C&C views provide means to abstract away direct hierarchy, direct connectivity, port names and types, and thus can crosscut the traditional boundaries of the implementation-oriented hierarchical decomposition of systems and sub-systems, and reflect the partial knowledge available to different stakeholders involved in a system’s design. As a primary application for C&C views we investigate the synthesis problem: given a C&C views specification, consisting of mandatory, alternative, and negative views, construct a concrete satisfying C&C model, if one exists. We show that the problem is NP-hard and solve it, in a bounded scope, using a reduction to SAT, via Alloy. We further extend the basic problem with support for library components, specification patterns, and architectural styles. The result of synthesis can be used for further exploration, simulation, and refinement of the C&C model or, as the complete, final model itself, for direct code generation. A prototype tool and an evaluation over four example systems with multiple specifications show promising results and suggest interesting future research directions towards a comprehensive development environment for the structure of component and connector designs.

💡 Research Summary

The paper introduces Component‑and‑Connector (C&C) views as a novel abstraction for describing the structural aspects of software and system architectures. Traditional C&C models require a fully specified hierarchy, explicit connections, and concrete port names and types, which makes early‑stage design cumbersome and limits the ability to capture partial knowledge held by different stakeholders. C&C views deliberately relax these requirements: they allow omission of hierarchy, direct connectivity, and port details, and they can be combined in three modalities—mandatory, alternative, and negative—to express, respectively, required elements, optional design choices, and prohibited configurations. By cross‑cutting traditional decomposition boundaries, views enable a more intuitive, stakeholder‑centric specification of architectural intent.

The central research problem is the synthesis of a concrete C&C model that satisfies a given set of views. Formally, the input consists of a finite universe of components, ports, and connectors together with a collection of view constraints. The output must be a model that (i) contains every element demanded by mandatory views, (ii) satisfies at least one alternative view for each alternative group, and (iii) violates none of the negative views. The authors prove that this synthesis problem is NP‑hard by reduction from SAT, establishing that any efficient algorithm would essentially solve Boolean satisfiability.

To obtain practical solutions, the authors adopt a bounded‑scope approach. They fix upper limits on the numbers of components, ports, and connectors, then encode the synthesis problem into Alloy, a declarative modeling language that translates specifications into propositional formulas for a SAT solver. In the encoding, each component, port, and connector becomes a Boolean atom; mandatory views generate conjunctive clauses, alternative views generate disjunctive clauses, and negative views generate negated literals. The resulting SAT instance is solved by an off‑the‑shelf SAT engine; a satisfying assignment is then decoded back into a concrete C&C model.

Beyond the basic synthesis engine, the paper extends the framework in three important directions:

1. Library Components – Pre‑defined reusable components are modeled as templates. During synthesis the solver may instantiate any library component that fits the constraints, promoting reuse and reducing the search space.

2. Specification Patterns – Common architectural patterns such as pipelines, broadcast, and pipe‑filter are captured as higher‑level constraints. By embedding these patterns, the synthesized model automatically respects well‑known structural idioms.

3. Architectural Styles – Style‑specific rules (e.g., layered architecture must be acyclic, client‑server must have a single server) are added as global constraints, ensuring that the final model conforms to a chosen style.

The authors implemented a prototype tool that integrates the Alloy encoder, the SAT solver, and a post‑processing step that visualizes the generated C&C diagram. They evaluated the approach on four case studies: an embedded control system, a data‑processing pipeline, a distributed service platform, and a robotic control framework. Each case involved multiple mandatory, alternative, and negative views, as well as library components, patterns, and style constraints. The experiments showed that, for models containing roughly 10–15 components, synthesis times ranged from 2 to 5 seconds on a standard workstation, demonstrating the feasibility of the approach for early‑stage design exploration. The generated models were subsequently used for simulation, incremental refinement, and direct code generation, illustrating the practical utility of the synthesis output.

The paper also discusses limitations. The bounded‑scope assumption means that scalability degrades sharply as the number of components grows, a typical characteristic of SAT‑based methods. Moreover, the current formulation does not handle dynamic connector creation, runtime reconfiguration, or sophisticated type systems, which are common in modern microservice and IoT architectures.

In conclusion, the work makes several contributions: (a) a flexible, stakeholder‑oriented view language for partial architectural specifications; (b) a formal proof of the synthesis problem’s computational hardness; (c) a concrete SAT‑based synthesis pipeline that can incorporate libraries, patterns, and style constraints; and (d) empirical evidence that the method can produce useful models for realistic systems. Future research directions proposed include incremental synthesis (updating an existing model when new views are added), optimization‑guided synthesis (minimizing component count, communication cost, or other metrics), integration with alternative SMT solvers (e.g., Z3) for richer theories, and extensions to capture dynamic or timed aspects of architectures. If these avenues are pursued, C&C view‑driven synthesis could become a cornerstone of model‑driven engineering, bridging the gap between high‑level architectural intent and concrete, verifiable system implementations.