【摘要】：在云存儲(chǔ)中,為了保護(hù)數(shù)據(jù)安全及用戶隱私,通常需對(duì)數(shù)據(jù)加密。與明文不同,加密后的數(shù)據(jù)往往難以操作,除了數(shù)據(jù)屬主之外,其他用戶無法訪問,這種限制嚴(yán)重影響了數(shù)據(jù)的可用性,數(shù)據(jù)的使用價(jià)值也大大降低。為了在保證數(shù)據(jù)安全及用戶隱私的同時(shí)提高數(shù)據(jù)的可用性,本文針對(duì)云存儲(chǔ)環(huán)境的特點(diǎn),深入研究了LSH(Locality Sensitive Hashing)相似性搜索方法,并根據(jù)云存儲(chǔ)海量數(shù)據(jù)的特點(diǎn),在現(xiàn)有LSH方法的基礎(chǔ)上對(duì)查詢過程中的第二階段加入偽相似對(duì)象剔除,提出了E-LSH(Efficient LSH)相似性搜索方法,在不影響準(zhǔn)確率的前提下有效提高了查詢效率,解決了加密造成的數(shù)據(jù)蔽塞。E-LSH方法采用索引與存儲(chǔ)分離的方式,將特征項(xiàng)索引與密文數(shù)據(jù)分開存儲(chǔ),對(duì)特征項(xiàng)進(jìn)行轉(zhuǎn)碼處理,用戶在搜索數(shù)據(jù)過程中無法接觸數(shù)據(jù),既保護(hù)了用戶的查詢隱私,又保證了數(shù)據(jù)的安全。為了適用云存儲(chǔ)海量數(shù)據(jù)搜索的需求,本文采用Map Reduce實(shí)現(xiàn)了E-LSH方法,并將其運(yùn)行于Hadoop分布式環(huán)境,實(shí)驗(yàn)結(jié)果表明E-LSH方法與Multi-Probe LSH方法相比加速比達(dá)到18.4%。此外,本文還設(shè)計(jì)了一個(gè)云存儲(chǔ)加密數(shù)據(jù)搜索系統(tǒng)方案,在保證數(shù)據(jù)安全的基礎(chǔ)上實(shí)現(xiàn)了數(shù)據(jù)的搜索與安全共享。系統(tǒng)支持關(guān)鍵詞和數(shù)據(jù)對(duì)象查詢,系統(tǒng)構(gòu)建者也可根據(jù)不同數(shù)據(jù)類型定制所需的系統(tǒng),既滿足了用戶的數(shù)據(jù)搜索與共享需求,又方便了系統(tǒng)構(gòu)建者和云存儲(chǔ)服務(wù)提供商。
[Abstract]:In cloud storage, data is usually encrypted to protect data security and user privacy. Different from plaintext, encrypted data is often difficult to operate, except for the main data, other users can not access. This restriction seriously affects the availability of data, and the use value of data is greatly reduced. In order to improve the availability of data while ensuring data security and user privacy, the similarity search method of LSH (Locality Sensitive Hashing) is studied in depth according to the characteristics of cloud storage environment, and according to the characteristics of cloud storage mass data, On the basis of existing LSH methods, pseudo-similar objects are removed in the second stage of the query process, and a E-LSH (Efficient LSH) similarity search method is proposed, which can effectively improve the query efficiency without affecting the accuracy of the query. The method of E-LSH uses the method of separating index and storage, stores the index of feature item and ciphertext data separately, transcodes the feature item, and the user can not contact the data in the process of searching data. It not only protects the user's query privacy, but also ensures the security of the data. In order to meet the requirement of cloud storage mass data search, this paper uses Map Reduce to implement the E-LSH method and runs it in the distributed environment of Hadoop. The experimental results show that the acceleration ratio of the E-LSH method compared with the Multi-Probe LSH method is 18.4. In addition, this paper also designs a cloud storage encryption data search system, which realizes data search and security sharing on the basis of ensuring data security. The system supports keyword and data object query, and the system builder can customize the system according to different data types, which not only meets the data search and sharing needs of users, but also facilitates the system builder and cloud storage service provider.
【學(xué)位授予單位】：國防科學(xué)技術(shù)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2013
【分類號(hào)】：TP333;TP309.7
，

本文編號(hào)：2414573