Date of Award

December 2016

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Electrical Engineering and Computer Science

Advisor(s)

C.Y. Roger Chen

Keywords

Data Processing, Debugging, Java Bytecode, Visualization

Subject Categories

Engineering

Abstract

We devise a widely applicable debugging approach to deal with the prevailing issue that bugs cannot be precisely reproduced in nondeterministic complex concurrent programs. A distinct efficient record-and-playback mechanism is designed to record all the internal states of execution including intermediate results by injecting our own bytecode, which does not affect the source code, and, through a two-step data processing mechanism, these data will be aggregated, structured and parallel processed for the purpose of replay in high fidelity while keeping the overhead at a satisfactory level. Docker and Git are employed to create a clean environment such that the execution will be undertaken repeatedly with a maximum precision of reproducing bugs. In our development, several other forefront technologies, such as MongoDB, Spark and Node.js are utilized and smoothly integrated for easy implementation. Altogether, we develop a system for Java Debugging Execution Visualization and Verification (JDevv), a debugging tool for Java although our debugging approach can apply to other languages as well. JDevv also offers an aggregated and interactive visualization for the ease of users’ code verification.

Access

Open Access

Download

Included in

Engineering Commons

Share

COinS