NOTES ON [Efficient and Cost-Effective Hybrid Congestion Control for HPC Interconnection Networks]

Efficient and Cost-Effective Hybrid Congestion Control for HPC Interconnection Networks 

Abstract—Interconnection networks are key components in high-performance computing (HPC) systems, their performance having a strong influence on the overall system one. However, at high load, congestion and its negative effects (e.g., Head-of-line blocking) threaten the performance of the network, and so the one of the entire system. Congestion control (CC) is crucial to ensure an efficient utilization of the interconnection network during congestion situations. As one major trend is to reduce the effective wiring in interconnection networks to reduce cost and power consumption, the network will operate very close to its capacity. Thus, congestion control becomes essential. Existing CC techniques can be divided into two general approaches. One is to throttle traffic injection at the sources that contribute to congestion, and the other is to isolate the congested traffic in specially designated resources. However, both approaches have different, but non-overlapping weaknesses: injection throttling techniques have a slow reaction against congestion, while isolating traffic in special resources may lead the system to run out of those resources. In this paper we propose EcoCC, a new Efficient and Cost-Effective CC technique, that combines injection throttling and congested-flow isolation to minimize their respective drawbacks and maximize overall system performance. This new strategy is suitable for current commercial switch architectures, where it could be implemented without requiring significant complexity. Experimental results, using simulations under synthetic and real trace-based traffic patterns, show that this technique improves by up to 55 percent over some of the most successful congestion control techniques.

关注的问题:互联网络【拥塞控制(CC)】

现有CC技术及存在问题(两种技术分别存在问题且相互不重叠):

  • 控制源节点流量 (反应时间长)
  • 隔离拥塞流量 (资源耗尽问题)

文章提出方案:EcoCC

  • 源控与隔离相结合的方案来提高系统性能
时间: 2024-11-09 14:18:55

NOTES ON [Efficient and Cost-Effective Hybrid Congestion Control for HPC Interconnection Networks]的相关文章

Android Change TCP Congestion Control

I The need of Change TCP Congestion Control in Android " TCP was originally designed for wired networks. Packet loss is considered to be the result of network congestion and the congestion window size is reduced dramatically as a precaution. However,

Network | TCP congestion control

拥塞控制算法:1. 加性增.乘性减:2. 慢启动:3. 对超时事件作出反应: 整体过程如下: 慢启动->到达阈值->加性增(窗口+1个MSS), 这个阶段叫拥塞避免(CA)->3个冗余ack丢包(事件)->阈值和窗口都缩小为一半(乘性减), 然后加性增(CA)->快速重传,并等待确认,这个阶段叫快速恢复,如果没有确认,就当作超时事件处理了: ->超时(事件)->快速重传->tcp会重新回为原始状态,进入慢启动: 在发生丢包和超时时,tcp都会执行快速重传:

论文解读《Understanding the Effective Receptive Field in Deep Convolutional Neural Networks》

感知野的概念尤为重要,对于理解和诊断CNN网络是否工作,其中一个神经元的感知野之外的图像并不会对神经元的值产生影响,所以去确保这个神经元覆盖的所有相关的图像区域是十分重要的:需要对输出图像的单个像素进行预测的任务,使每一个输出像素具有一个比较大的感知野是十分重要的,在做预测试时,每一个关键的信息就不会被遗漏. 增大感知野的方法: 理论上可以通过搭建更多的层的网络实现感知域的线性增加,靠着卷积过滤器的增加: 也可以使用下采样的方法,池化,增加感知域,目前通常都结合了这两种技术: 作者发现 并不是所

Autodesk 2014 M&E Devday is coming

年度的欧特克传媒娱乐业的开发者大会又将很快与小伙伴们见面了,我们会在拉斯维加斯,东京,上海,北京四个地方分别举办.相比往年,今年会晚一些,也让我们有更多的时间去准备.以下是英文的详细介绍,如果只关注中国的话,小伙伴们只要看红色字体标注部分就可以了,任何问题,可以留言或者邮件给我 [email protected] .. Autodesk Media and Entertainment DevDays December 2014 – January 2015 Come see us in acti

Congestion Avoidance in TCP

Congestion Avoidance in TCP Consequence of lack of congestion control When a popular resource is shared without regulation the result is always over-utilization With the introduction of TCP in 1983, users can write networking applications that requir

(转)Awesome Courses

Awesome Courses  Introduction There is a lot of hidden treasure lying within university pages scattered across the internet. This list is an attempt to bring to light those awesome courses which make their high-quality material i.e. assignments, lect

Heterogeneous Self-Organizing Network for Access and Backhaul

This application discloses methods for creating self-organizing networks implemented on heterogeneous mesh networks. The self-organizing networks can include a computing cloud component coupled to the heterogeneous mesh network. In the methods and co

Linux Ethernet Bonding Driver HOWTO 英文原版

Linux Ethernet Bonding Driver HOWTO Latest update: 12 November 2007 Initial release : Thomas Davis <tadavis at lbl.gov> Corrections, HA extensions : 2000/10/03-15 : - Willy Tarreau <willy at meta-x.org> - Constantine Gavrilov <const-g at xp

Google C++ 代码规范

Google C++ Style Guide Table of Contents Header Files Self-contained Headers The #define Guard Forward Declarations Inline Functions Names and Order of Includes Scoping Namespaces Unnamed Namespaces and Static Variables Nonmember, Static Member, and