Creating Real-time Applications in the Java Programming Language

摘要

The portability promise of the Java programming language - expressed in the now-famous "Write Once, Run Anywhere" slogan - accounts for much of Java's popularity among developers of web-based and enterprise applications. It has also sparked a growing interest in Java among developers of real-time and embedded systems. The portability benefits for the embedded market are clear: developers can reuse application code across product lines and deploy it on new hardware or software platforms, without having to recertify the entire system - provided, of course, the Java runtime environment itself has been certified. That said, both the standard Java language specification and the Java Virtual Machine (VM) specification are incompatible with true real-time programming, due to their limitations in predictability, garbage collection, and other capabilities essential to ensuring timely completion of tasks. To address these shortcomings and enhance Java's suitability for real-time and embedded systems, a team of experts from Sun, QNX, Nortel, and other organizations developed the "Real-time Specification for Java," or RTSJ. The team, called the Real-Time for Java Expert Group, submitted their Java Specification Request as part of the Java Community Process. The specification was ratified in 2000, and several vendors now offer commercial implementations of the RTSJ, making real-time programming in Java a reality. In this paper, we introduce the guiding principles that the expert group employed to create the specification, discuss the resulting extensions to the Java specifications, and illustrate some of the programming interfaces and patterns for real-time programming in Java.