Tarry’s Algorithm and Consistency in Tree-like Distributed
Systems

Mykola Terletskyi, Grygoriy Zholtkevych

Анотація


In distributed systems, efficient and predictable network traversal is critical for coor-
dination, broadcasting, and maintaining data consistency between nodes. This paper
examines the behavior of Tarry’s traversal algorithm in tree-like undirected networks
under varying workloads, with a specific focus on consistency. Using simulated environ-
ments with 10, 100, and 1000 nodes, we initiated multiple transaction waves (T = 2,
5, 10) to evaluate the algorithm’s ability to support consistent data propagation. Our
results show that while Tarry’s algorithm, due to its cycle-free and deterministic nature,
ensures that all nodes are visited without duplication or omission, it alone is insuffi-
cient to guarantee data consistency. Although it provides a reliable traversal foundation,
maintaining complete consistency in tree-like distributed systems requires additional syn-
chronization or coordination mechanisms to ensure consistency. These findings highlight
the strengths and limitations of the algorithm in supporting consistent state propagation
in distributed environments.


Повний текст:

PDF (English)

Посилання


1. https://ieeexplore.ieee.org/document/6133253

2. https://api.semanticscholar.org/CorpusID:60440625

3. https://mitpress.mit.edu/9780262037662/distributed-algorithms/

4. https://dl.acm.org/doi/10.5555/2721747

5. https://dl.acm.org/doi/10.1145/359545.359563

6. https://iopscience.iop.org/article/10.1088/0967-1846/2/4/005

7. https://www.researchgate.net/publication/2377958_Robot_Navigation_in_Unknown_Terrains_Introductory_Survey_of_Non-Heuristic_Algorithms

8. https://dl.acm.org/doi/abs/10.5555/2821576

9. https://diestel-graph-theory.com/

10. https://dl.acm.org/doi/abs/10.5555/3153775

11. https://dl.acm.org/doi/10.1145/3133933

12. https://link.springer.com/chapter/10.1007/978-3-031-79018-8_9




DOI: http://dx.doi.org/10.30970/vam.2025.34.13635

Посилання

  • Поки немає зовнішніх посилань.