IOSIG: Declarative I/O-Stream Properties Using Pragmas

Author:

Gholami MasoudORCID,Schintke Florian

Abstract

AbstractNon-functional properties of IO-streams are typically not specified and passed to the used middleware or operating system. Knowing properties such as the expected access pattern, reliability, and visibility for data would allow a better storage resource selection by the infrastructure and thus could improve overall performance. With pragma annotations, we let developers declare the intended data access for their file descriptors and automatically redirect the corresponding input/output to the most appropriate storage device of the user’s system. Our automatic resource selection lets applications leverage system storage capabilities better, often delivers improved performance (in some cases by orders of magnitude), and relieves uninformed users from having to make the best choice manually.

Funder

Konrad-Zuse-Zentrum für Informationstechnik

Publisher

Springer Science and Business Media LLC

Subject

General Earth and Planetary Sciences,General Environmental Science

Reference46 articles.

1. (2017) symlink(2) – Linux manual page. https://man7.org/linux/man-pages/man2/symlink.2.html. Accessed 01 May 2022

2. Afify GM, El Bastawissy A, Hegazy OM (2016) Identifying hot/cold data in main-memory database using frequent item set mining. Int J Enhanc Res Manag Comput Appl 5(3):35

3. Alter J et al (2019) SSD failures in the field: symptoms, causes, and prediction models. In: Proc. SC. ACM, pp 75:1–75:14 https://doi.org/10.1145/3295500.3356172

4. Axboe J fio – flexible IO tester. https://git.kernel.dk/?p=fio.git. Accessed 01.2021

5. Bjørling M et al (2013) Linux block IO: introducing multi-queue SSD access on multi-core systems. In: 6th SYSTOR. ACM, pp 22:1–22:10 https://doi.org/10.1145/2485732.2485740

Cited by 1 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. Proactive Resource Management to Optimize Distributed Workflow Executions;2023 IEEE International Conference on Big Data (BigData);2023-12-15

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3