Where it Stands: Actual vs Symbolic Spaces
In order to be processed, objects must be “at hand”, i.e activities must know where to find them and then fetch them. That will be straightforward for local objects but may entail architectural constraints when targeted objects cannot be fetched directly (if at all), and therefore instantly.
Location and Architectural Constraints
System Modelling is meant to encompass both actual contexts and symbolic representations under a single roof, and legacy and new applications along a single time scale.
The first rationale behind object locations is that if an activity has to be performed as a whole, it must have complete control over its resources, which means instant and secure access to targeted objects given the activity location.
A dual rationale is to enforce the integrity of objects targeted by simultaneous activities: contrary to a single location, multiple ones will bear specific architectural constraints: authorization, confidentiality, synchronization, etc.
Hence, as far as system architectures are concerned, two aspects must be clarified as soon as possible:
- The actual location of objects and the agents using them.
- The relative location of symbolic representations with cross dependencies. Their actual location will be set depending on system architecture.