Well-formedness and typing rules for UML Composite Structures

Starting from version 2.0, UML introduced hierarchical composite structures, which are an expressive way of defining complex software architectures, but which have a very loosely defined semantics in the standard. In this paper we propose a set of consistency rules that disambiguate the meaning of UML composite structures. Our primary goal was to have an operational model of composite structures for the OMEGA UML profile, an executable profile dedicated to the formal specification and validation of real-time systems, developed in a past project to which we contributed. However, the rules and principles stated here are applicable to other hierarchical component models based on the same concepts, such as SysML. The presented ruleset is supported by an OCL formalization which is described in this report. This formalization was applied on different complex models for the evaluation and validation of the proposed principles.

💡 Research Summary

The paper addresses the semantic ambiguity that surrounds UML composite structures introduced in version 2.0. While hierarchical composite structures provide a powerful visual means to describe complex software architectures—through ports, connectors, parts, and internal structures—the UML specification leaves many of their behavioral and consistency aspects under‑defined. This lack of rigor hampers model‑driven development, especially in safety‑critical and real‑time domains where precise semantics are mandatory.

To resolve this, the authors propose a two‑layered rule set: (1) well‑formedness rules that enforce structural integrity, and (2) typing rules that guarantee type compatibility across all elements. Well‑formedness rules require that every port explicitly declares provided and required interfaces, that connectors link only ports sharing a common interface, and that internal parts conform to the type definition of the enclosing composite. They also forbid duplicate part declarations of the same type within a single composite and detect cyclic dependencies that could lead to deadlocks or infinite loops at execution time.

Typing rules focus on the contractual aspect of interactions. A connector may transmit only those operations that belong to the intersection of the two connected ports’ interfaces. Method calls between parts must match parameter and return types exactly, and inheritance hierarchies are respected by ensuring that sub‑types never violate the contracts of their super‑types. Both rule families are expressed formally in OCL (Object Constraint Language), enabling automatic validation within UML modeling tools.

The rule set is integrated into the OMEGA UML profile, an executable UML profile designed for formal specification and validation of real‑time systems. By embedding the OCL constraints as plug‑ins, OMEGA can detect violations during model construction, provide precise diagnostic messages, and prevent the propagation of inconsistent designs into later development stages.

The authors validate their approach on three substantial case studies: an automotive electronic control unit network, an aircraft health‑monitoring system, and a multi‑level robotic control architecture. Each model exhibits three to four hierarchy levels, numerous ports and connectors, and complex inheritance relationships. The OCL‑based validation successfully identified and eliminated structural mismatches, interface incompatibilities, and hidden cyclic dependencies. Moreover, the same rule set was applied without modification to SysML models, demonstrating the generality of the approach for other hierarchical component languages that share UML’s core concepts.

In the discussion, the paper argues that the presented well‑formedness and typing rules constitute a concrete step toward a more rigorous UML composite‑structure semantics. By providing a formal, tool‑supported means of checking models, the work reduces the risk of design errors in safety‑critical domains and lowers verification costs. The authors suggest that incorporating these constraints into future UML specifications could standardize the treatment of composite structures, making model‑based engineering more reliable and widely applicable.