MCSD-100

MCSD100 is a dataset for object storage performance testing, based on the MSCOCO and semEval2017 datasets used in previous studies (Sherlock in OSS). The dataset is designed to evaluate the performance of the Open-Stack Swift cloud storage system, specifically its Erasure Coding policy and fragmentation capabilities. The dataset includes 25 files each of text, mp3, jpeg, and mp4 formats, representing real-world scenarios. By analyzing the behavior of these files under different EC schemes, the researchers gain valuable insights into OpenStack Swift's storage mechanisms. The dataset is openly available for reproducibility and further studies. The findings highlight OpenStack Swift's efficiency in handling different file types and sizes, making it suitable for enterprise data management, multimedia content platforms, image hosting, long-term data preservation, and collaborative content creation.

MCSD100是一款面向对象存储性能测试的数据集，其依托此前研究中使用的微软COCO（MSCOCO）数据集与2017年国际语义评测（SemEval2017）数据集构建（即开源存储领域的Sherlock基准）。本数据集旨在评估OpenStack Swift云存储系统的性能，具体针对其纠删码（Erasure Coding）策略与分片能力。数据集包含文本、MP3、JPEG、MP4四种格式的文件各25份，覆盖真实应用场景。通过分析这些文件在不同纠删码方案下的存储行为，研究人员可深入洞察OpenStack Swift的存储机制。本数据集公开可获取，以支持研究复现与后续拓展研究。相关研究结果验证了OpenStack Swift在处理不同类型与大小文件时的高效性，使其可适用于企业数据管理、多媒体内容平台、图床服务、长期数据留存以及协同内容创作等诸多场景。

搜集汇总

背景与挑战

背景概述

MCSD-100是一个用于对象存储性能测试的数据集，基于MSCOCO和semEval2017数据集构建，专门评估OpenStack Swift云存储系统的纠删码策略和分片能力。它包含25个文件，覆盖文本、mp3、jpeg和mp4格式，模拟真实场景，公开可用，适用于企业数据管理、多媒体平台和长期数据保存等应用。