Adaptive Hybrid Secondary Storage Systems for Cloud Computing Scenarios

云计算场景的自适应混合二级存储系统

• 批准号: 259252678
• 负责人: Professor Dr.-Ing. Gero Mühl
• 金额: --
• 依托单位: Lehrstuhl für Architektur von Anwendungssystemen
• 依托单位国家: 德国
• 项目类别: Research Grants
• 财政年份: 2014
• 资助国家: 德国
• 起止时间: 2013-12-31 至 2017-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/259252678?language=en
• 关键词: Adaptive; Hybrid; Secondary; Storage; Systems

The project aims at improving the performance of secondary storage systems in cloud computing scenarios. This is achieved by using a holistic system architecture in combination with dynamically adapting the data placement and organization in order to better exploit the heterogeneous characteristics of different kinds of storage devices. The placement is continuously adjusted according to current access patterns and the specific technological characteristics of the storage devices, which differ in performance, capacity, and price. Additionally, the optimization also takes the requirements of the cloud applications and of the platform operator into account. By using a holistic system architecture, which is located on the file system layer and which integrates the functionality of the usual storage stack, available information can be exploited more directly, more flexible data organizations can be used, and more fine-grained adaptations can be executed. The developed solutions are validated by extensive simulations and evaluated in a realistic testbed running on an HPC cluster.

该项目旨在提高云计算场景下二级存储系统的性能。这是通过使用整体系统架构与动态适应数据放置和组织相结合来实现的，以便更好地利用不同类型存储设备的异构性。根据当前的访问模式和存储设备的特定技术特征不断调整布局，这些存储设备在性能、容量和价格上存在差异。此外，优化还考虑了云应用和平台运营商的要求。通过使用位于文件系统层并且集成了通常存储堆栈的功能的整体系统架构，可以更直接地利用可用的信息，可以使用更灵活的数据组织，并且可以执行更细粒度的适配。所开发的解决方案通过大量的模拟验证，并在运行在HPC集群上的真实试验台上进行了评估。

项目成果

An Adaptive SSD Cache Architecture Simultaneously Using Multiple Caches

• DOI: 10.1109/nas.2019.8834724
• 发表时间: 2019-08
• 期刊: 2019 IEEE International Conference on Networking, Architecture and Storage (NAS)
• 作者: Nikolaus Jeremic;Helge Parzyjegla;Gero Mühl

Improving random write performance in heterogeneous erasure-coded drive arrays by offloading code block requests

通过卸载代码块请求来提高异构纠删码驱动器阵列中的随机写入性能

• DOI: 10.1145/2695664.2695696
• 发表时间: 2015
• 期刊: Proceedings of the 30th Annual ACM Symposium on Applied Computing
• 作者: Jeremic;Nikolaus;Helge Parzyjegla;Gero Mühl
• 通讯作者: Gero Mühl