Where it Happens: Actual vs Symbolic Locations
Actual activities are meant to change business contexts, either directly on object states, or indirectly on agent expectations; as a consequence, their execution will necessarily have to be mapped to objects and agents locations. While symbolic activities can be performed wherever computing resources are at hand, symbolic dependencies may also have to take into account distributed locations.
Since system architectures will be directly affected by accesses to objects and synchronization, constraints on address spaces and time scalesof activities should be defined up front, especially for distributed activities involving different clocks.
Accesses, Authorizations, Synchronization
From a business perspective locations are set by the origin of events, if any, affecting the activity, ie the location of the agents at the source. From a system perspective the architecture will will have to map the locations of entry points with those of targeted resources, and introduce some communication mechanism if time is set by different clocks. Hence, if architectural constraints are to be properly identified at requirements level then mapped into archetypes at design level, activities must be decomposed so that execution units may be set in single address spaces, to be run under single clocks.
That decomposition is essential if identification and authorizations are to be properly managed: access rights being assessed at system entry points, they must subsequently be managed within the corresponding location.