- 學位論文
有效更新及檢驗為基礎之隱私保護多關鍵字排序搜尋
Privacy-Preserving Multi-keyword Ranked Search with Verifying and Updating Efficiently
摘要
由於雲端服務的崛起，越來越多個人與企業選擇將龐大且重要的 資料放在雲端上，交由強力的雲端服務提供者協助管理，如此一來可 以減少自行空間管理以及硬體上的支出，但是不管是公開雲或是私有 雲，將資料交由第三方管理都有安全性上的威脅，多數資料擁有者會 選擇加密上傳，卻衍生了在密文搜尋不易的問題。最近，Cao 等人發 表 MRSE 技術，提供密文下的多關鍵字搜尋以及關聯性排序結果，根 據關鍵字在文件中的重要性做到文件排名，若有文件更新的情況出現， 則需要重建搜尋索引，而在現實中，資料更新的頻率將會造成 MRSE 的效率低落。 在本篇論文中，我們利用修改反向索引來改善更新效率不佳的問 題，藉由雲端串流文件的元數據給使用者，讓使用者可以計算文件關 聯性結果，而在有威脅的環境下，我們的設計也不會洩漏有關資料額 外的訊息，使用者還可以驗證索引的正確性防止雲端有錯誤或是作弊 的情況。在此設計下，使用者雖負擔了一些計算量，但仍可以在快速 的時間內得到驗證過的多關鍵字搜尋及關聯性排序結果。
並列摘要
Because of the cloud service springing up quickly, more and more in- dividuals and enterprise choose to store the massive and significant data in cloud. The cloud service providers offer the powerful service to help manage data, in this way the users can reduce the expendicure on storage management and hardware. However, the user will confront the threat on security because we delegate the management to the third-party. Most user will encrypt the data before outsourcing, but this behavior give rise to the problem about the searching under the encrypted data. In recent years, Cao et al. [1] propose the MRSE technique which can provide multi-keyword ranked search. If in the case of updating document, this MRSE need to rebuild the whole searchable index. In the reality, the frequency of the updating documents will lower the MRSE efficiency. In this work, we make efforts in improving the updating efficiency. The cloud stream the index metadata to the user, and the user de- crypt the metadata and get the ranked result. In the threatened environment, our design will not leak any privacy about the document. Moreover, the user even can verify the data integrity to prevent the cloud cheating or storage er- ror accident. In our framework, the user cooperate with the cloud and burden some caculation, but we can still get the verified multi-keyword ranked result in the short time.