The per-process view of naming and remote execution

The authors focus on naming issues in remote execution. They use an abstraction that provides a per-process view of naming rather than the common per-site view. This allows a process to create an arbitrary view of naming where each process has its own root node to which name spaces from different parts of the distributed system are attached. Ordered merges of contexts are also used for further control in configuring a name space. The result is a flexible naming environment, on which a powerful facility for both transparent and nontransparent remote execution was constructed