Coordinated Atomic actions (CA actions) provide a scheme for coordinating complex concurrent activities and supporting error recovery between multiple interacting objects in a distributed object-oriented system. In this paper we show how CA actions can be applied to two different areas: safety-critical systems and fault-tolerant parallel systems. We have used a Production Cell case study to show how we can use CA actions to control a safety-critical system, where safety requirements play a fundamental role, and we have used an example based on the GAMMA paradigm to demonstrate how CA actions can be used to add fault tolerance to a parallel computation model. We discuss how CA actions provide these systems with dependability features, and describe our Java framework for constructing CA actions. Keywords: coordinated atomic actions, dependability, Java, object-oriented systems, safety-critical systems 1 Introduction The purpose of the research described in this paper is to demonstrate ...

In the late 80's new applications such as CAD, hyper-media, and programming environments stressed to the limit existing database technology [2]. To meet the demands of these new domains OODBs were developed. There was, and still is a great diversity in their architecture and functionality. Furthermore, this functionality continues to evolve [14, 22]. However, early on one common feature emerged and that is they all used some distinct service that provided stable storage for objects or pages, referred to as object servers and page servers respectively [13]. This paper describes problems facing the continued development of object stores and addresses them in the design of ObServer2. The design takes a new approach to the construction of an object server that allows it to meet changing needs. Also, it presents original contributions in the design of concurrency control and recovery algorithms. New applications such as CAD, hyper-media, and programming environments stress existing ...