HDT: A Heuristic Dynamic Threshold Algorithm to Avoid Reprioritization of LEDBAT
-
摘要: 近年来,随着通信技术和网络传输能力的大幅度提升,应用需求呈现多元化的增长态势(视频会议、在线游戏等交互式应用要求低时延、低抖动,而软件更新等应用则要求高吞吐).为满足时延不敏感的数据传输并保证高效的瓶颈带宽利用率,低优先级拥塞控制算法(如LEDBAT(low extra delay background transport))受到广泛关注.该类算法能在链路空闲时占用未被使用的带宽,而在链路负载较高时释放占用的带宽以保证时延敏感数据的传输.然而,当中间路由器部署主动队列管理算法时,低优先级拥塞控制算法存在优先级反转问题,即链路高负载时无法释放占用的带宽,使其退化为普通拥塞控制算法.为解决该问题,针对LEDBAT中的固定时延阈值造成的优先级反转,提出启发式的动态阈值调整算法,其在运行时动态搜索最优的动态时延阈值,确保LEDBAT与主动队列管理算法共存时仍能保持低优先级特性,同时不降低链路的利用率.为验证算法的有效性,在网络模拟NS2中建立了不同网络场景并对算法进行大量的评估.实验结果表明:与已有低优先拥塞控制算法相比,新算法能够有效解决优先级反转的问题,同时保证链路的带宽利用率.Abstract: In recent years, with the significant increase in communication technologies and network transmission capabilities, the requirements of applications have shown a diversified growth trend (such as video conferencing and online games with low latency, low delay jitter and software updates with high throughputs). In order to meet delay-insensitive data transmission and ensure efficient bottleneck bandwidth utilization, low priority congestion control (LPCC) algorithms such as LEDBAT(low extra delay background transport) have received increasing attention. These algorithms can occupy available bandwidth when the link is idle, and release the occupied bandwidth when the link load is high to ensure the transmission of delay-sensitive data. However, when AQM(active queue management) is deployed on the router, the reprioritization problem of LPCC occurs. When the link is under heavy load, the occupied bandwidth cannot be released by LPCC, which degenerates into an ordinary congestion control algorithm. To solve the problem, this paper proposes a heuristic dynamic threshold adjustment algorithm, called HDT which can dynamically search for the optimal dynamic delay threshold, ensuring it maintains low priority while coexisting with AQM without reducing link utilization. We establish different network scenarios in the network simulator NS2 to verify the effectiveness of the proposed algorithm. The results show that HDT can effectively solve the problem of reprioritization while ensuring the bandwidth utilization of the link.
-
-
期刊类型引用(7)
1. 李曼文,张月琴,张晨威,张泽华. 异质图嵌入的地理不敏感时空兴趣点推荐方法. 计算机科学与探索. 2024(03): 755-767 . 百度学术
2. 金柯君,于洪涛,吴翼腾,李邵梅,张建朋,郑洪浩. 改进的基于奇异值分解的图卷积网络防御方法. 计算机应用. 2023(05): 1511-1517 . 百度学术
3. 王小红,刘琴. 基于深度迁移的有向加权网络节点重叠检测. 计算机仿真. 2023(09): 492-496 . 百度学术
4. 金柯君,于洪涛,李邵梅,张建朋. 基于注意力机制的图卷积网络防御方法. 信息工程大学学报. 2023(06): 718-724 . 百度学术
5. 杨旭华,王磊,叶蕾,张端,周艳波,龙海霞. 基于节点相似性和网络嵌入的复杂网络社区发现算法. 计算机科学. 2022(03): 121-128 . 百度学术
6. 刘志鑫,张泽华,张杰. 基于多层次多视角的图注意力Top-N推荐方法. 计算机科学. 2021(04): 104-110 . 百度学术
7. 陈晋音,黄国瀚,张敦杰,张旭鸿,纪守领. 一种面向图神经网络的图重构防御方法. 计算机研究与发展. 2021(05): 1075-1091 . 本站查看
其他类型引用(4)
计量
- 文章访问数: 1012
- HTML全文浏览量: 3
- PDF下载量: 145
- 被引次数: 11