Scholarlink

<p>File providers in large-scale distributed systems, such as modern data centers, face the critical challenge of efficiently sharing vast volumes of data as the number of users continues to grow, all while maintaining consistent performance. Distributed Data Centers play a pivotal role in addressing this challenge by enabling resource replication across multiple servers, thereby optimizing bandwidth utilization and ensuring scalability. This paper introduces a large-scale dynamic Distributed Data Center model grounded in BitTorrent and inspired by the fairness principles of the cake-cutting game, tailored specifically for large-scale environments. The proposed model leverages BitTorrent’s sharing mechanisms and incorporating the equitable resource allocation of the cake-cutting game to enhance overall system efficiency. Its performance is rigorously evaluated through simulations in three distinct scenarios, focusing on critical metrics relevant to distributed systems. The results underscore the superior efficiency of the proposed approach, particularly in large-scale settings. Specifically, the model achieves a 99.11% improvement in delay for responding to nodes’ initial requests for each piece compared to alternative approaches. Additionally, it delivers a 24.04% reduction in peer-to-peer response time, further solidifying its suitability for large-scale distributed applications.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A dynamic game theoretic resource allocation utilization in high-load torrent based large-scale distributed data centers

  • 期刊论文

  • 发表日期：

    2026年2月28日

查看全文
  • Pardis Kallesari,
  • Peyman Bayat,
  • Gholamhossein Ekbatanifard

摘要

File providers in large-scale distributed systems, such as modern data centers, face the critical challenge of efficiently sharing vast volumes of data as the number of users continues to grow, all while maintaining consistent performance. Distributed Data Centers play a pivotal role in addressing this challenge by enabling resource replication across multiple servers, thereby optimizing bandwidth utilization and ensuring scalability. This paper introduces a large-scale dynamic Distributed Data Center model grounded in BitTorrent and inspired by the fairness principles of the cake-cutting game, tailored specifically for large-scale environments. The proposed model leverages BitTorrent’s sharing mechanisms and incorporating the equitable resource allocation of the cake-cutting game to enhance overall system efficiency. Its performance is rigorously evaluated through simulations in three distinct scenarios, focusing on critical metrics relevant to distributed systems. The results underscore the superior efficiency of the proposed approach, particularly in large-scale settings. Specifically, the model achieves a 99.11% improvement in delay for responding to nodes’ initial requests for each piece compared to alternative approaches. Additionally, it delivers a 24.04% reduction in peer-to-peer response time, further solidifying its suitability for large-scale distributed applications.