A framework for location transparency in distributed systems

A common issue of distributed software systems is the allocation of the logical software components of a system to physical nodes. In particular there is often a need for "location transparency", that is, the ability to map a single specification to a number of different physical environments without requiring changes to the code. We describe a formal framework that can be used to achieve location transparency. This framework, based on a formalized concept of "computing environments", is defined for distributed systems specified using the ROOM structural modeling language. However it can easily be extended to other object based architectural languages.