Narek Galstyan

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2024-4

January 12, 2024

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2024/EECS-2024-4.pdf

This report reviews bug catalogs and debugging systems designed for distributed systems. It tries to find common pat- terns in distributed systems bugs, highlights the characteris- tics necessary in a debugging system to identify these bugs in distributed systems, and proposes the Application-Integrated Record-Replay (aiRR) system for addressing classes of these bugs. aiRR is designed specifically for distributed systems. aiRR integrates the recording into the distributed system and lever- ages this integration to reduce the overhead of recording in the application. To have low overhead, our approach avoids reducing application-level concurrency and avoids recording application-level data that is not necessary for replay.

Advisors: Scott Shenker and Sylvia Ratnasamy

---

BibTeX citation:

@mastersthesis{Galstyan:EECS-2024-4,
    Author= {Galstyan, Narek},
    Title= {Application-Integrated Record-Replay of Distributed Systems},
    School= {EECS Department, University of California, Berkeley},
    Year= {2024},
    Month= {Jan},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2024/EECS-2024-4.html},
    Number= {UCB/EECS-2024-4},
    Abstract= {This report reviews bug catalogs and debugging systems designed for distributed systems. It tries to find common pat- terns in distributed systems bugs, highlights the characteris- tics necessary in a debugging system to identify these bugs in distributed systems, and proposes the Application-Integrated Record-Replay (aiRR) system for addressing classes of these bugs.
aiRR is designed specifically for distributed systems. aiRR integrates the recording into the distributed system and lever- ages this integration to reduce the overhead of recording in the application. To have low overhead, our approach avoids reducing application-level concurrency and avoids recording application-level data that is not necessary for replay.},
}

EndNote citation:

%0 Thesis
%A Galstyan, Narek 
%T Application-Integrated Record-Replay of Distributed Systems
%I EECS Department, University of California, Berkeley
%D 2024
%8 January 12
%@ UCB/EECS-2024-4
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2024/EECS-2024-4.html
%F Galstyan:EECS-2024-4