Files
bachelor-thesis/data/abstract.tex
2026-05-27 23:38:36 +08:00

30 lines
4.3 KiB
TeX
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
% !TEX root = ../bachelor-thesis.tex
% 中英文摘要和关键字
\begin{abstract}
跨域覆盖网络需要为用户提供稳定的低延迟、高带宽和低丢包传输服务的同时降低运营成本。传统方案通常依赖专线链路保证服务质量,但专线根据带宽成本计算,且成本较高。已有的成本优化方法主要分为两类,链路调度方法尝试在公网质量较好时使用公网分流,而端到端方法尝试通过调整发送策略或网络编码应对链路质量下降,但实际测量表明,公网链路质量下降时段往往与用户流量高峰重合,难以有效削减专线峰值成本。已有端到端方法能够缓解丢包影响,但未区分不同链路片段的质量差异,容易在质量良好的片段上引入额外带宽开销,造成带宽浪费。
针对上述问题本文提出一种面向跨域公网覆盖网络的分段链路质量修复方法。该方法在全部使用公网连接的前提下仅对低质量链路片段启用交织XOR分组编码。文本对链路建立三状态丢包信道模型根据解码端上报的丢包统计动态选择编码参数同时设计基于PI控制器的输出速率控制机制缓解FEC解码按组恢复导致的突发输出问题。
本文使用Rust语言实现了原型系统并在模拟跨域低质量链路的实验环境中进行验证。实验结果表明本文方法能够正确识别低质量链路片段并启用FEC修复在无丢包链路片段上保持普通转发。在0至~\SI{2}{\percent} 链路丢包率范围内与直接转发方案相比本文方法最高达到约3.6倍吞吐提升,验证了分段链路质量修复方法的有效性。
% 关键词用“英文逗号”分隔,输出时会自动处理为正确的分隔符
\thusetup{
keywords = {覆盖网络, 前向纠错编码, 公网链路优化, 链路质量修复},
}
\end{abstract}
\begin{abstract*}
Cross-domain overlay networks need to provide users with stable low-latency, high-bandwidth, and low-loss transmission services while reducing operational costs. Traditional approaches usually rely on dedicated links to guarantee service quality, but dedicated links are often charged by bandwidth and incur high costs. Existing cost optimization methods mainly fall into two categories. Link scheduling methods try to offload traffic to public Internet links when their quality is good, while end-to-end methods try to cope with link quality degradation by adjusting sending strategies or applying network coding. However, real-world measurements show that quality degradation of public Internet links often overlaps with user traffic peaks, making it difficult to effectively reduce the peak bandwidth cost of dedicated links. Existing end-to-end methods can alleviate packet loss, but they do not distinguish quality differences among individual link segments. As a result, they may introduce unnecessary bandwidth overhead on high-quality segments and waste bandwidth resources.
To address these problems, this thesis proposes a segment-level link quality repair method for cross-domain public-Internet overlay networks. Under an all-public-Internet interconnection setting, the method enables interleaved XOR block coding only on low-quality link segments. This thesis builds a three-state packet loss channel model for each link and dynamically selects coding parameters according to packet loss statistics reported by the decoder. It also designs an output rate control mechanism based on a PI controller to mitigate the bursty output caused by group-based FEC recovery.
This thesis implements a prototype system in Rust and evaluates it in an experimental environment that emulates low-quality cross-domain links. The results show that the proposed method correctly identifies low-quality link segments and enables FEC repair on them, while keeping normal forwarding on loss-free segments. With link loss rates from 0 to \SI{2}{\percent}, the proposed method achieves up to about 3.6x throughput improvement compared with direct forwarding, demonstrating the effectiveness of segment-level link quality repair.
% Use comma as separator when inputting
\thusetup{
keywords* = {overlay network, forward error correction, public Internet link optimization, link quality repair},
}
\end{abstract*}