Overloading of Names in Distributed Systems

Service request mechanisms in distributed systems have developed from a connectionbased paradigm, in which users specify services by naming the hardware by which the service is to be rendered. Implicit in this form of naming is the specification of the characteristics of the desired service. In other words, names of physical devices have been overloaded with the specification of type of service that the hardware offers. This overloading is undesirable, because it forces users to be aware of the implicit connection between hardware name and type of service. and because it binds services to particular hardware very early in the service execution process. We present an example of connection-based overloading (based on the 4.2BSD UNIX printer queue utilities Ipr,lpq, and Iprm). and propose a service-oriented naming mechanism.

[1]  A. Retrospective,et al.  The UNIX Time-sharing System , 1977 .

[2]  Richard W. Watson,et al.  Identifiers (Naming) in Distributed Systems , 1980, Advanced Course: Distributed Systems.