The Research And Implementation On Secure Index Technology Based On Full-text Searching System

With the rapid development of Internet, information searching technology has been aimportant tool in our life. We put forward the concept of secure index, because we mustprotect users’ information in some scene. That means we protect users’ information byusing encryption and hash, and we can do some searching on the premise of that theinformation won’t leak out.According to the problems currently proposed two secure index modes. The first oneis based on inverted index. It has taken inverted index as its structure, the text is encryptedin the stages of index and query, and it takes a second encryption when writing into theindex files, so that can inherited the advantages of speed and accurately from invertedindex, but for the chosen-plaintext attack and frequency statistics and other means ofattack is not yet effective defense. Because of the inverted index lack of security, anothermode has been proposed: the mode based on trapdoor one-way function. Combining thetrapdoor one-way function with pseudo-random function, can well against attacks such aschosen plaintext, thus can protect the users’ information more secure. But thecomputational complexity, inefficient, large space occupied, and a certain probability offalse positives are the disadvantages of this mode.Describe the principle and security of two modes, and achieve two secure indexingsystems: S-Lucene and BF-Index. S-Lucene is based on Lucene, which is a open sourcesearch engine; BF-Index using Bloom Filter as its index storage structure. Compare theperformance of two systems in all aspects through experiments, such as index buildingtime, searching time and accuracy. Analyze advantages and disadvantages of two systemsby experimental data, and propose a clear future direction.