Empirical evaluation of two main-stream RTSJ implementations

  • Authors:
  • James Mc Enery;David Hickey;Menouer Boubekeur

  • Affiliations:
  • National University of Ireland, Cork;National University of Ireland, Cork;National University of Ireland, Cork

  • Venue:
  • JTRES '07 Proceedings of the 5th international workshop on Java technologies for real-time and embedded systems
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

The Real-Time Specification for Java (RTSJ) is a set of interfaces and behavioural specifications that allow for realtime programming in Java. It is designed to support both hard and soft real-time applications. Among its major features are: scheduling properties, asynchronous event handling, asynchronous transfer of control and enhanced memory management. Since the release of the RTSJ, many implementations have been developed conformant to this specification. Their performance on certain aspects of the RTSJ can vary greatly. For real-time programmers it is important to highlight these variations so that the most suitable implementation is chosen for the application being developed. This paper evaluates the performance of two commercial releases according to specific features of the RTSJ. The two implementations we choose are Java RTS (Java Real-Time System) from Sun-Microsystems and JamaicaVM from Aicas which are both widely used in academia and industry. Efficiency and predictability, which are important considerations when designing a real-time application, are considered. Specifically, the features analysed are: memory allocation, thread management, synchronisation and asynchronous event handling. This evaluation will benefit the development and improvement of future RTSJ implementations.