A microscopic view of bursts, buffer contention, and loss in data centers
A microscopic view of bursts, buffer contention, and loss in data centers
复制标题
DOI：
10.1145/3517745.3561430
复制
发表时间：
2022-10
期刊：
影响因子：
--
通讯作者：
Ehab Ghabashneh;Yimeng Zhao;C. Lumezanu;N. Spring;S. Sundaresan;Sanjay G. Rao
中科院分区：
文献类型：
--
作者：
Ehab Ghabashneh;Yimeng Zhao;C. Lumezanu;N. Spring;S. Sundaresan;Sanjay G. Rao
管理低损耗的数据中心网络需要了解短（毫秒）时间尺度的流量动态，尤其是流量的突发性，以及突发流量在多大程度上争夺交换机缓冲区资源。然而，在这样的时间间隔内监控流量是一个大规模的挑战。我们做出了两项贡献。首先，我们介绍 Millisampler，这是一种部署在所有 Meta 主机上的轻量级流量表征工具。 Millisampler 采用以主机为中心的视角来收集数据，它具有可扩展性，并且允许将流量模式与传输层统计数据相关联。此外，跨机架中的服务器同时收集 Millisampler 数据可以分析同步流量如何在机架缓冲区中交互。特别是，我们研究了当多个突发同时到达动态共享机架缓冲区时发生的争用。其次，我们提出了数据中心规模的争用分析，包括对突发性、争用和丢失的独特联合分析。我们的结果表明 (i) 争用特征在区域内和区域内差异很大，并且受到服务布局的影响； (ii) 争用在短期内变化很大； (iii) 突发可能会遇到一些争用； (iv) 较高的争用不一定会导致更多的损失，并且与工作负载和突发属性的相互作用很重要。我们讨论对数据中心设计的影响，包括服务布局、缓冲区共享算法和拥塞控制。
Managing data center networks with low loss requires understanding traffic dynamics at short (millisecond) time-scales, especially the burstiness of traffic, and to what extent bursts contend for switch buffer resources. Yet, monitoring traffic over such intervals is a challenge at scale. We make two contributions. First, we present Millisampler, a lightweight traffic characterization tool deployed across all Meta hosts. Millisampler takes a host-centric perspective to data collection, which is scalable and allows for correlating traffic patterns with transport layer statistics. Further, simultaneous collection of Millisampler data across servers in a rack enables analysis of how synchronized traffic interacts in rack buffers. In particular, we study contention, which occurs when multiple bursts arrive simultaneously at the dynamically shared rack buffer. Second, we present a data-center-scale analysis of contention, including a unique joint analysis of burstiness, contention, and loss. Our results show (i) contention characteristics vary widely across and within a region and is influenced by service placement; (ii) contention varies significantly over short time-scales; (iii) bursts are likely to encounter some contention; and (iv) higher contention need not lead to more loss, and the interplay with workload and burst properties matters. We discuss implications for data center design including service placement, buffer sharing algorithms and congestion control.