Thesis (Ph. D.)--University of Washington, 2005. / Vita. Includes bibliographical references (p. 135-146).

Thesis (M.S.)--Oregon State University, 2007. / Printout. Includes bibliographical references (leaves 34-36). Also available on the World Wide Web.

Thesis (M.S.)--Oregon State University, 2008. / Printout. Includes bibliographical references (leaves 51-55). Also available on the World Wide Web.

Thesis (Ph. D.)--University of Washington, 2000. / Vita. Includes bibliographical references (p. 253-264).

This thesis investigates event-based behavioural model debugging in Linda. A study is presented of the Linda parallel programming paradigm, its amenability to debugging, and a model for debugging Linda programs using Milner's CCS. In support of the construction of expected behaviour models, a Linda program specification language is proposed. A behaviour recognition engine that is based on such specifications is also discussed. It is shown that Linda's distinctive characteristics make it amenable to debugging without the usual problems associated with paraUel debuggers. Furthermore, it is shown that a behavioural model debugger, based on the proposed specification language, effectively exploits the debugging opportunity. The ideas developed in the thesis are demonstrated in an experimental Modula-2 Linda system.

Thesis (M.S.)--Oregon State University, 2010. / Printout. Includes bibliographical references (leaves 48-51). Also available on the World Wide Web.

published_or_final_version / abstract / toc / Educational Psychology / Master / Master of Social Sciences

published_or_final_version / abstract / Computer Science / Doctoral / Doctor of Philosophy

Short time-to-bug localization is extremely important for any 24x7 service-oriented application. To this end, we introduce a new debugging paradigm called live debugging. There are two goals that any live debugging infrastructure must meet: Firstly, it must offer real-time insight for bug diagnosis and localization, which is paramount when errors happen in user-facing applications. Secondly, live debugging should not impact user-facing performance for normal events. In large distributed applications, bugs which impact only a small percentage of users are common. In such scenarios, debugging a small part of the application should not impact the entire system. With the above-stated goals in mind, this thesis presents a framework called Parikshan, which leverages user-space containers (OpenVZ) to launch application instances for the express purpose of live debugging. Parikshan is driven by a live-cloning process, which generates a replica (called debug container) of production services, cloned from a production container which continues to provide the real output to the user. The debug container provides a sandbox environment, for safe execution of monitoring/debugging done by the users without any perturbation to the execution environment. As a part of this framework, we have designed customized-network proxies, which replicate inputs from clients to both the production and test-container, as well safely discard all outputs. Together the network duplicator, and the debug container ensure both compute and network isolation of the debugging environment. We believe that this piece of work provides the first of its kind practical real-time debugging of large multi-tier and cloud applications, without requiring any application downtime, and minimal performance impact.

Thesis (M.S.)--Oregon State University, 2008. / Printout. Includes bibliographical references (leaves 35-37). Also available on the World Wide Web.