Implementing an ADA Kernel on NEBULA.
Abstract
NEBULA and Ada are rapidly approaching use. These have been designed to achieve similar goals, to improve the quality, timeliness, and cost of the real-time systems used in the Department of Defense. NEBULA is intended to provide a common architecture for these systems in applications that range from microprocessors to mainframes. Ada is designed, among other things, to ease concurrent programming. This report reviews the concurrency features of Ada, examines the aspects of the NEBULA architecture that are more significant for the implementation of concurrent programs, suggests a method for reducing the tasking mechanisms of Ada to a few simple kernel operations, and evaluates the NEBULA architecture in terms of this method and these operations. It is found that NEBULA supports admirably the control structures of Ada, but its Memory Management system is not very suitable. Entry calls and Accept statements are found to require about 40 machine instructions each. A simple rendez-vous requires the execution of over 100 instructions.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1983
- Accession Number
- ADA132745
Entities
People
- Giorgio P. Ingargiola
Organizations
- Temple University