EUVIMEDEUVIMED entdecken
Crossref

Optimizing Execution Time in Distribute Networks an Empirical Study of Node Configuration and Concurrency Levels

R. Saravanan

REST Journal on Data Analytics and Artificial Intelligence · 01.03.2025 · 4 (1) · 726

DOI: 10.46632/jdaai/4/1/95 ↗

Kurz gefasst

Fachartikel in der Zeitschrift REST Journal on Data Analytics and Artificial Intelligence, online erschienen am 01.03.2025. Verfasst von R. Saravanan. Die Arbeit nennt 23 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This research explores the complex interactions between system parameters and performance metrics in a distributed system, with an emphasis on concurrency control. Distributed systems support modern computing, supporting large-scale web services, transaction processing, and data-intensive applications. Effectively managing concurrent operations across multiple nodes while ensuring consistency and performance is challenging. This study analyzes how the number of nodes, concurrency levels, network latency, and execution time affect system behavior. Using a dataset of 100 observations with 2 to 49 nodes and concurrency levels ranging from 1 to 31, performance patterns are examined using descriptive statistics and advanced regression models. The average network latency is 166.43 ms, with execution times averaging 2065.56 ms, showing significant variation across configurations. Both linear regression and random forest regression were used to predict execution time based on system parameters. Linear regression achieved R² values of 0.8587 (training) and 0.7936 (test), while random forest regression performed better with R² values of 0.9886 and 0.8543. Correlation analysis showed a strong positive relationship between synchronization and execution time (0.84) and a negative relationship with the number of nodes (-0.41), highlighting the impact of synchronization overhead. These findings highlight important trade-offs in distributed system design, showing that the increased parallelism through additional nodes must be balanced against communication overhead and network latency. The results provide valuable guidance for architects and developers aiming to improve performance, efficiently manage resources, and design scalable, concurrent systems with minimal execution delays.

Zitieren

APA

Saravanan, R. (2025). Optimizing Execution Time in Distribute Networks an Empirical Study of Node Configuration and Concurrency Levels. REST Journal on Data Analytics and Artificial Intelligence, 4(1), 726. https://doi.org/10.46632/jdaai/4/1/95

Vancouver

Saravanan R. Optimizing Execution Time in Distribute Networks an Empirical Study of Node Configuration and Concurrency Levels. REST Journal on Data Analytics and Artificial Intelligence. 2025;4(1):726. doi:10.46632/jdaai/4/1/95

BibTeX
@article{Saravanan2025,
  author = {Saravanan, R.},
  title = {Optimizing Execution Time in Distribute Networks an Empirical Study of Node Configuration and Concurrency Levels},
  journal = {REST Journal on Data Analytics and Artificial Intelligence},
  year = {2025},
  volume = {4},
  number = {1},
  pages = {726},
  doi = {10.46632/jdaai/4/1/95}
}
Kostenlos starten →