In response to ever-increasing requirements, next-generation embedded systems will typically consist of several physically dispersed uniprocessor or multiprocessor nodes. In order to harness the potential of such architectures, a definition of distributed, fault-tolerant Ada must exist. Because the language standard does not meaningfully address distribution or fault tolerance, an extension of the underlying runtime support environment is both appropriate and cost effective. This proposal outlines initial baselines for the necessary models of distribution, fault tolerance and dynamic reconfiguration, and proposes to refine these models into a preliminary systems architecture design.
Keywords: Runtime Systems, Distribution, Fault Tolerance, Ada