To facilitate reuse, cooperation and parallelization of constraint solvers, the CCOPS project has defined a new execution model for distributed-memory systems. It represents a constraint solver as a set of closure operators acting on constraints and variable domains together with a strategy for applying them. Variable domains and constraints are local structures and accordingly, an operator represents a sequential computation. In this manner, a system design emerges where reuse of existing solvers amounts to their inclusion as operators, cooperation is a mixture of local interleaving and control-parallelism, and data-parallelism occurs when operators are coordinated between many processes. The three paradigms can be dynamically combined by strategies expressed in this model, realized a new system called BS-Solve (bulk synchronous solver environment or BSS for short).