Current network transmission mechanisms frequently lack the precision and flexibility required to accommodate the increasingly diverse and personalized quality-of-service (QoS) expectations of modern users. These mechanisms often adopt fixed congestion control strategies that are not sufficiently adaptable to dynamic network conditions or varying user demands. To address this challenge, we introduce a novel transmission control mechanism, Ditto. This mechanism is designed with an efficient application-layer interface and incorporates a user satisfaction metric function that enables dynamic selection of congestion control algorithms tailored to user-specific demands. Additionally, Ditto optimizes packet loss handling through active retransmission techniques and is implemented within the open-source XQUIC framework. Experimental results demonstrate that Ditto can flexibly adjust transmission strategies in response to the dynamic evolution of user needs. In scenarios involving delivery rate sensitivity, latency sensitivity, and packet loss sensitivity, Ditto exhibits strong adaptability, effectively fulfilling user application requirements, and significantly enhancing both network performance and user experience.

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

Ditto: A Flexible Transmission Control Mechanism for Diverse User Demands and Network Conditions

  • Qingnan Wang,
  • Jinnuo Du,
  • Fangzhou Chen,
  • Danfeng Shan

摘要

Current network transmission mechanisms frequently lack the precision and flexibility required to accommodate the increasingly diverse and personalized quality-of-service (QoS) expectations of modern users. These mechanisms often adopt fixed congestion control strategies that are not sufficiently adaptable to dynamic network conditions or varying user demands. To address this challenge, we introduce a novel transmission control mechanism, Ditto. This mechanism is designed with an efficient application-layer interface and incorporates a user satisfaction metric function that enables dynamic selection of congestion control algorithms tailored to user-specific demands. Additionally, Ditto optimizes packet loss handling through active retransmission techniques and is implemented within the open-source XQUIC framework. Experimental results demonstrate that Ditto can flexibly adjust transmission strategies in response to the dynamic evolution of user needs. In scenarios involving delivery rate sensitivity, latency sensitivity, and packet loss sensitivity, Ditto exhibits strong adaptability, effectively fulfilling user application requirements, and significantly enhancing both network performance and user experience.