Single-user capabilities in interprocess communication
Stephen C. Russell · ACM SIGOPS Operating Systems Review · 1988
Many distributed operating systems use interprocess communication (IPC) and data transfer schemes which are optimised for a particular architecture: the workstation/server model. In this model, client processes have full control over data transfer. This paper argues that server controlled data transfer is better suited to systems with a hierarchy of servers distributed across a collection of machines. The system proposed integrates a set of IPC primitives, for the exchange of control information, and a capability system for secure data transfer.