JBInsTrace: A tracer of Java and JRE classes at basic-block granularity by dynamically instrumenting bytecode

Abstract

a b s t r a c tUnderstanding what happens during the runtime of a Java program is difficult. Tracking runtime flow can bring valuable information for program understanding and behavior analysis. Polymorphism, thread concurrency or even simple facts like the number of method invocations and the number of executed bytecodes are valuable information to track, but are difficult to compute outside the Java Virtual Machine (JVM) on running programs. In this paper, we present JBInsTrace, a new tool that instruments and traces Java bytecode. It produces static information about source code and a very fine grained trace of Java software execution, combining them to allow detailed analysis of the runtime. Our tool differs from others because it does not only trace program classes but also JRE classes, and does so at basic block level, without altering the JVM and without statically modifying class files. We explain JBInsTrace design, focused towards efficiency, which results in reasonable performance penalty.

Distributed API Protocol Mining

“…ISpecMiner and ISpecMiner+ are nearly the same: 1) both tools employ the Java agent [17] technique to instrument client programs; 2) both tools mine API protocols based on the online approach proposed by Chen et al [25]. The only difference is that ISpecMiner is a local miner, while ISpecMiner+ is a distributed API protocol miner.…”

“…Once an instrumented program is ran with test cases generated automatically or manually, the embedded codes will output information of method calls sequentially. Whatever instrumentation techniques and tools can be used in our framework (such as ASM [15], BCEL [16], Java agent [17] and Javassist [18]- [19]), provided that the format of output PETs can satisfy the requirement of NSpecMiner. After that, the client sends gathered PETs to the server via NCM, one method call after another.…”

Visualizing trace of Java collection APIs by dynamic bytecode instrumentation

“…Work in [27] addresses the problem from a data mining perspective and uses association rules mining technique to find frequently used APIs. Caserta et al [11] developed an instrumentation tool that traces Java program. The tool captures static information and at the same time performs analysis by capturing runtime data.…”

“…The work presents performance results of the tool on five Java programs. However, the work in [11] has no suitable visual interface and case studies are not listed.…”

Mining Universal Specification Based on Probabilistic Model

“…Many approaches and frameworks exist to instrument Java applications statically or dynamically. We adopt Java agent technique, which is a service provided by Java since 1.5 [18]. Java agents can instrument classes at bytecode level.…”