Research Of Ciphertext Inquery On Chinese Document In Sharing Outsourced Database Mode

Under Outsourcing database mode, It's not safe to store plain data at the untrusted remote database servers because of the sensitivity of the document data, that is, the document data should be encrypted before storage. After Encryption, the encrypted data loses the characteristics of document data and can not be operated by common method supported by database servers. while most of the database operations are based by searching, so how to use'Select'on ciphertext becomes a hot topic.In the text, the author designs two kinds of Ciphertext document retrieval subsystems: the document retrieval subsystem based on word for word concordance and the document retrieval subsystem based on keywords index, the document retrieval subsystem based on word for word concordance keeps the characteristics of the plaintext by keeping the position and the existing characters in the plaintext, the whole search process can be divided into two steps: first , get all the records that match the input condition by users; second, decrypt the records generated by step 1 ,and search again on the decrypted records; the document retrieval subsystem based on keywords index searches by keeping the existing keywords characteristic in the plaintext, the system will split all the keywords existing in the plaintext before encrypted and build keywords index at the database, so it can get all the records that contain the keywords. besides, user can use one or more words without knowing too much about the articles existing in the Database.According to the Experimental results, the document retrieval subsystem based on word for word concordance has a highly Precision and better query performance than other methods referenced in the text while the document retrieval subsystem based on keywords index has a better performance than the document retrieval subsystem based on word for word concordance as long as the split program is accurate.