ReLoaDing Performance: A Locality-Based Strategy for Rapid Reads in Encrypted Key-Value Systems

NCU Institutional Repository > 資訊電機學院 > 資訊工程研究所 > 博碩士論文 > Item 987654321/96293

jsp.display-item.identifier=請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/96293

题名:	ReLoaDing Performance: A Locality-Based Strategy for Rapid Reads in Encrypted Key-Value Systems
作者:	洪季絜;Hung, Chi-Chieh
贡献者:	資訊工程學系
关键词:	鍵值儲存;快閃記憶體;資料安全;資料加密;Key-value store;flash memory;data security;data encryption
日期:	2024-12-03
上传时间:	2025-04-09 17:37:42 (UTC+8)
出版者:	國立中央大學
摘要:	在鍵值儲存系統中，為了確保資料的安全性，常採用壓縮與加密技術來處理鍵值對。然而，將多個鍵值對進行打包以提升壓縮比和加密效率的方式，卻在僅需存取部分資料時，容易導致讀取延遲的增加。隨著時間推移，資料的存取模式不斷在改變，傳統的打包方式逐漸無法適應實際的資料區域性，導致系統需反覆解密與解壓縮大量資料，進一步造成效能損耗。本研究提出了一種名為 ReLoaD 的創新區域性重新打包策略，該策略針對高頻存取的鍵值對進行動態重新打包，不僅降低了解密與解壓縮的成本，還有效縮短了讀取延遲，全面提升了加密鍵值儲存系統的效能。實驗結果證實，ReLoaD 在確保資料安全性的同時，大幅改善了讀取效率，為打造更高效且具適應性的安全資料儲存方案提供了新方向。;In key-value store systems, data security is often prioritized through compression and encryption of stored key-value pairs. However, packaging multiple key-value pairs together for improved compression ratios and encryption efficiency leads to significant read latency when only a subset of data is required. As data access patterns evolve over time, traditional storage packages no longer align with actual data locality, resulting in performance overhead due to the repeated decryption and decompression of large data units. This paper introduces ReLoaD, a novel locality-remodeling packing strategy designed to dynamically repack frequently accessed key-value pairs. By intelligently restructuring data locality, ReLoaD reduces the decryption and decompression burdens, minimizing read latency and enhancing overall performance in encrypted key-value store applications. Experimental results demonstrate that ReLoaD achieves substantial improvements in read efficiency without compromising security, paving the way for more responsive and adaptable secure data storage solutions.
显示于类别:	[資訊工程研究所] 博碩士論文

文件中的档案:

档案	描述	大小	格式	浏览次数
index.html		0Kb	HTML	52	检视/开启

在NCUIR中所有的数据项都受到原著作权保护.

社群 sharing

数据加载中.....