The Research Of Quering Method Over Encryption Character Data In Database Based On B*Tree Clustered Index

In the DAS model, the user’s data was stored in the database servers which wereprovided by third party. The third party was responsible for the query, management and hadadministrator privileges of database to access all the data at any time. Under thiscircumstance, sensitive data in the database was not confidentiality to the third party. Inorder to ensure the security of sensitive data, encryption methods were adopted to encryptthe data.Howerer,that increased the difficulty of query over encrypted data.In order to improve query performance over the encrypted string data, this paperpresented a method of query over encrypted string data based on the B*tree clustered indexin database. Integer index characteristic value was generated by sensitive string data throughIndex Characteristic Function and was stored in the new index field. At the same time,B*tree clustered index was created on the new index field. A cluster made records which hadthe same or related the characteristic values store in a data block (or a plurality of datablocks in connection with a linked list), which reduce the access time with reducing the diskinput and output. Creating B*tree index on index field could quickly locate the address ofthe data in order to query the data, thus avoiding a full table scaning.When querying over sensitive string data, sensitive string data was converted to indexcharacteristic value which matched with sensitive string data, B*tree clustered index canquickly locate the address of the data in order to query the data, thus avoiding a full tablescan.In order to improve the fuzzy query performance over encrypted string data, this paperincrease two data in the leaf nodes: one was a pointer array; the other was to identify thedomain. The pointer array was the t address of the leaf nodes which meet certain rules. Thedomain was used to identify whether the right pointer is equal to The first pointer in pointerarray. Fuzzy query can complete based on B*tree clustered index with the pointer array andright pointer in the leaf node help, thus avoiding scanning all leaf node, to improve queryspeed.The result of experiment showed that this method had better query efficiency.