WCET analysis of Java bytecode featuring common execution environments

Abstract

We present a novel tool for statically determining the Worst Case Execution Time (WCET) of Java Bytecode-based programs called Tool for Execution Time Analysis of Java bytecode (TetaJ). This tool differentiates itself from existing tools by separating the individual constituents of the execution environment into independent components. The prime benefit is that it can be used for execution environments featuring common embedded processors and software implementations of the JVM. TetaJ employs a model checking approach for statically determining WCET where the Java program, the JVM, and the hardware are modelled as Networks of Timed Automata (NTA) and given as input to the state-of-the-art UPPAAL model checking tool. The tool is evaluated through a case study based on the classic text-book example of a hard realtime control system in a mine pump. The system is hosted on an execution environment featuring an interpretationbased JVM, called Hardware near Virtual Machine (HVM) that runs on an Atmel AVR ATmega2560 processor.

Symbolic execution and timed automata model checking for timing analysis of Java real-time systems

“…In recent years, several analysis tools for SCJ have been put forward including WCA [22], SARTS [23], TetaJ [24], and TETASARTS [25]. Characteristic for these is that they rely on reconstruction of the control-flow graph (CFG), which characterizes all possible execution paths of the program.…”

“…The NTA model can readily be used for estimating WCET and BCET but also generalizes to verification of properties expressible in timed computation tree logic (TCTL). In contrast to previous tools [24,25], which provide limited feedback, our technique can also generate witness traces that expose the reported behavior, useful for debugging and program understanding.…”

“…Many model-based timing analysis tools such as META-MOC [7] for executables, and WCA [22], TetaJ [24], SARTS [23], and TETASARTS [25] for Java rely on processing the program to build a CFG and enumerating the execution paths over the CFG. The paths are then translated to a modeling formalism, e.g., timed automata, and timing analysis is then formulated as a reachability problem using an appropriate logic such as TCTL.…”

“…Inevitably, considering all execution paths characterized by the CFG as in [7,22,24,25] affects the analysis time, but it may also severely over-approximate the temporal behavior.…”

“…For example, this will more precisely capture the effect of a shared pipeline and cache when tasks are scheduled. The potential benefits of this approach come at the expense of a bigger state space, and currently only small, although realistic, applications, such as the Minepump control system [1,24], are tractable for analysis.…”