Comparing data synchronization in Ada 9X and Orca

Protected object types are one of three major extensions to Ada 83 proposed by Ada 9X. This language feature is intended for light-weight data synchronization between tasks. The Orca parallel programming language has a very similar construct, the shared data-object, with which we have over five years of experience, both in usage and implementation. This paper compares protected objects and shared data-objects, with regard to design, usage, and implementation. Keywords: synchronization, parallel languages, distributed systems, language design. 1. INTRODUCTION In [1] we proposed a new model for communication between processes that run on distributed-memory machines. This model allows processes to share data structures that are encapsulated in shared dataobjects , which can be accessed through operations defined by an abstract data type. One of the key ideas in the model is that data synchronization is implicit, by executing all operations on shared data-objects indivisibly. During the..