【作者】 王楠；

【导师】 骆婷；

【作者基本信息】 华中科技大学 ， 信息安全， 2016， 硕士

【摘要】 大数据时代的到来,传统关系型数据库在存取速率等方面越来越无法满足海量数据处理的需求,NoSQL数据库应运而生。与关系型数据库相比,NoSQL数据库在安全方面存在不足。HBase分布式数据库是一种安全机制比较完善的NoSQL数据库,但HBase数据库中的数据文件却是以明文形式存储的,这便形成一定的安全漏洞。为HBase数据库中的数据文件进行加密处理,并为HBase数据库构建密文安全二级索引,使得HBase密文数据库具有条件查询能力是非常具有意义的。根据HBase分布式数据库中列数据的数据类型的不同可以分别研究数值型密文检索和文本型密文检索。对于HBase分布式数据库中数值型数据的密文检索,提出一种基于改进型安全R树的检索策略。在为HBase的数据表的数据构建R树安全索引时,会利用R树的动态特性根据数据表构建索引需要的列值来将RowKey行按照设定的加密策略存入到索引表的恰当位置。同时采用将HBase的数据表DataTable相对应的索引表IndexTable存入HBase中来确保对索引表使用与维护。这种方式能够很好的支持针对列的等值检索与范围检索。对于HBase分布式数据库中文本型数据的密文检索,提出一种基于安全倒排索引的检索策略。在构建文本型密文索引时,将HBase的常用于构建二级索引的coprocessor框架与开源搜索引擎工具包Lucene进行有机结合,将HBase中Column Family/Q ualifer限定的列规范为文本内容之后进行分词加密,同时采用将Lucene中的相关文件等信息保存在HBase数据库中来保证数据的完整性和可用性。这种方式可以很好地支持针对列的等值查询与子串查询。设计实现了基于HBase的密文检索原型系统,实现了用户管理模块、加解密模块、密钥管理模块、密文索引生成与检索模块等,并对系统进行了实验测试。HBase安全索引充分利用了现有HBase快速检索能力,目前的研究成果对密文数据查询多样性的支持仍有一定的局限,该方向仍具有极大的研究空间与研究意义。更多还原

【Abstract】 The arrival of the era of big data, the traditional relational database in the rate of access increasingly cannot meet the demand for massive data processing. The NoSQL database arises at the historic moment. Compared with the relational database, the NoSQL database is insufficient in the security aspect.HBase database is a kind of security mechanism in NoSQL database, but its file is stored in the data file in plain text form, this will form a certain security vulnerabilities.It is very meaningful to to encrypt the data files in the HBase database, and construct the two level index for the HBase database, which can increase the conditional query ability of the HBase database. Based on the data type of the column data in the HBase distributed database, we research the numerical ciper retrieval and text ciper retrieval.A ciper retrieval strategy based on improved R tree is proposed for the numerical data of the HBase in the distributed database. In building R tree index for the Data Table data in HBase Database,alse grounds the column values of constructing index in DataTable to store the row identified by rowkey to the proper position of index table in the fix encryption strategy by using the dynamic characteristics of the R- tree.. and the IndexTable corresponding to the Data Table in HBase datatable also store in HBase to keep the index table usage and maintenance.In the way,it can support the equivalent search and range retrieval for the column. A retrieval strategy based on secure inverted index structure is proposed for the text data in the HBase distributed database.When constructing text ciper index,we combine the coprocessor framework,which is commonly used in constructing the secondary index in HBase database and the open source search engine toolkit Lucene organically,and make the column specified by Column Family/Q ualifier standard into text content,and then participle、encrypt,and make the relevant documents and other information of the Lucene also store into HBase database to ensure the integrity and availability of the data.in this way,it can support the equivalent query and substring query for HBase column very well.Design and implement the prototype system based on HBase database,while implements the module of user management,the module of encryption and decryption,the module of key management,the module of the ciper text index’s generation and retrieval and so on,we also test the system.HBase security index makes full use of the existing HBase fast retrieval ability, the present research results on the ciphertext data query diversity support still has certain limitations, the direction still has great research space and the significance of the research.更多还原