SHF: EAGER: HI-HDFS - Holistic I/O optimizations for the Hadoop distributed filesystem

SHF：EAGER：HI-HDFS - Hadoop 分布式文件系统的整体 I/O 优化

• 批准号: 1747447
• 负责人: Spyros Blanas
• 金额: $ 15万
• 依托单位: Ohio State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-09-01 至 2018-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1747447&HistoricalAwards=false
• 关键词: SHF; EAGER; HI; HDFS; Holistic

File systems and their outdated POSIX "byte stream" interface suffer from an impedance mismatch with the versatile I/O requirements of today's applications. Specifically, the I/O path from the application to the raw storage device is becoming longer and it involves the interplay of intricate software and hardware components. This produces complex aggregate I/O patterns that application developers (often subject matter experts with limited knowledge of how massive concurrency creates I/O bottlenecks) cannot optimize based on intuition alone. File systems that tout their high scalability, such as the Hadoop distributed file system, largely do so by limiting applications to sequential access patterns. The question of whether one can accelerate the I/O performance of the Hadoop distributed file system for analytical applications with complex data models that cannot readily serialize data contiguously for fast sequential access remains open. This project seeks to address this question and build HI-HDFS -- a framework that automatically collects and manages semantically richer I/O metadata to guide object placement in the Hadoop distributed file system. The HI-HDFS framework synthesizes the I/O activity across software components throughout the datacenter in a navigable graph structure to identify application-agnostic motifs in I/O activity. A novel I/O forecasting technique identifies and ameliorates bottlenecks at large scale by inspecting I/O activity from small-scale runs. Overall, the HI-HDFS framework challenges the I/O optimization mantra that manual data placement is the cornerstone of I/O performance and paves the way towards next-generation object-centric storage systems for high-performance computers. The efficacy of this automated approach will be examined on a complex data processing workload from the domain of emergency response which exhibits I/O patterns that are characteristic of modern analytical applications. The broader impacts of this work are expected to include open-source prototype implementations as well as pedagogical impact on a cloud computing course for both Computer Science and Data Analytics undergraduate majors at Ohio State.

文件系统及其过时的 POSIX“字节流”接口与当今应用程序的多功能 I/O 要求存在阻抗不匹配的问题。具体来说，从应用程序到原始存储设备的 I/O 路径变得越来越长，并且涉及复杂的软件和硬件组件的相互作用。这会产生复杂的聚合 I/O 模式，应用程序开发人员（通常是对大规模并发如何造成 I/O 瓶颈了解有限的主题专家）无法仅凭直觉进行优化。宣扬高可扩展性的文件系统（例如 Hadoop 分布式文件系统）很大程度上是通过将应用程序限制为顺序访问模式来实现的。对于具有复杂数据模型的分析应用程序是否可以加速 Hadoop 分布式文件系统的 I/O 性能这一问题仍然悬而未决，因为这些复杂数据模型无法轻松地连续序列化数据以进行快速顺序访问。该项目旨在解决这个问题并构建 HI-HDFS——一个自动收集和管理语义上更丰富的 I/O 元数据以指导 Hadoop 分布式文件系统中的对象放置的框架。 HI-HDFS 框架以可导航的图形结构综合整个数据中心内软件组件的 I/O 活动，以识别 I/O 活动中与应用程序无关的主题。一种新颖的 I/O 预测技术通过检查小规模运行中的 I/O 活动来识别并改善大规模瓶颈。总体而言，HI-HDFS 框架挑战了 I/O 优化的信条，即手动数据放置是 I/O 性能的基石，并为高性能计算机的下一代以对象为中心的存储系统铺平了道路。这种自动化方法的有效性将在紧急响应领域的复杂数据处理工作负载上进行检验，该工作负载表现出现代分析应用程序特征的 I/O 模式。这项工作的更广泛影响预计将包括开源原型实现以及对俄亥俄州立大学计算机科学和数据分析本科专业云计算课程的教学影响。

项目成果

ATP: Directed Graph Embedding with Asymmetric Transitivity Preservation

• DOI: 10.1609/aaai.v33i01.3301265
• 发表时间: 2018-11
• 作者: Jiankai Sun-;Bortik Bandyopadhyay;Armin Bashizade;Jiongqian Liang;P. Sadayappan;S. Parthasarathy

ArrayBridge: Interweaving Declarative Array Processing in SciDB with Imperative HDF5-Based Programs

ArrayBridge：将 SciDB 中的声明性数组处理与基于 HDF5 的命令式程序交织在一起

• DOI: 10.1109/icde.2018.00092
• 发表时间: 2018
• 期刊: IEEE 34th International Conference on Data Engineering (ICDE
• 作者: Xing, Haoyuan;Floratos, Sofoklis;Blanas, Spyros;Byna, Suren;Prabhat, M.;Wu, Kesheng;Brown, Paul
• 通讯作者: Brown, Paul

ApproxJoin: Approximate Distributed Joins

• DOI: 10.1145/3267809.3267834
• 发表时间: 2018-10
• 期刊: Proceedings of the ACM Symposium on Cloud Computing
• 作者: D. Quoc;Istemi Ekin Akkus;Pramod Bhatotia;Spyros Blanas;Ruichuan Chen;C. Fetzer;T. Strufe

Evaluating Scalability Bottlenecks by Workload Extrapolation

• DOI: 10.1109/mascots.2018.00039
• 发表时间: 2018-09
• 期刊: 2018 IEEE 26th International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS)
• 作者: Rong Shi;Yifan Gan;Yang Wang

Characterizing I/O optimization opportunities for array-centric applications on HDFS

• DOI: 10.1109/hpec.2018.8547529
• 发表时间: 2018-09
• 期刊: 2018 IEEE High Performance extreme Computing Conference (HPEC)
• 作者: Donghe Kang;Vedang Patel;Kalyan Khandrika;Spyros Blanas;Yang Wang;S. Parthasarathy