KR101476039B1 - Method for encrypting database and method for real-time search thereof - Google Patents

Abstract

A method for encrypting a database is disclosed. To build encrypted data in a database server, a method for encrypting a database in a database client terminal generates an index by constructing a bloom filter for element units of a plaintext database, and generates an encrypted tuple by encrypting tuple units for the plaintext database.

Description

METHOD FOR ENCRYPTING DATABASE AND METHOD FOR REAL-TIME SEARCH THEREOF "

The present invention relates to a database encryption method and its real-time retrieval method.

Generally, a method for encrypting information stored in a database is performed for each unit constituting a database, and a table-based encryption method for encrypting an entire table, an attribute-based encryption method for encrypting a column unit, a tuple tuple) unit encryption method, and an element-based encryption method in which each element is encrypted.

The table-based and attribute-based methods are used for database backup, and real-time retrieval is not possible, and the element-based method is too much for encrypted data and can not be used for other retrieval except for the same value retrieval. On the other hand, since the tuple-based encryption method encrypts data in tuples, insertion and deletion in units of tuples can be performed, and a separate index for attributes in each tuple is required for real-time retrieval.

The elements constituting the database are classified into sensitive information and non-sensitive information. Non-sensitive information does not require encryption, but sensitive information needs to be encrypted. However, there is a problem that the encrypted database is difficult for the user to retrieve in real time.

It is an object of the present invention to provide a database encryption method and a real-time retrieval method thereof, which provide sufficient security for sensitive information stored in a database and enable real-time retrieval as well as insertion and deletion of data.

According to an embodiment of the present invention, there is provided a database encryption method for establishing encrypted data in a database server, the method comprising the steps of: ≪ / RTI > And encrypting the tuple unit of the plain text database to generate an encrypted tuple.

In one embodiment of the present invention, the encryption method may further include transmitting the Bloom filter and the encrypted tuple to a database server.

In one embodiment of the present invention, the encryption method may further include the step of the database server constructing an encrypted database using the Bloom filter and the encrypted tuple.

In an embodiment of the present invention, the database server may store the encrypted tuple and the Bloom filter as individual fields in a table.

In one embodiment of the present invention, the step of generating an index includes: extracting a keyword for an element unit of the plain text database; Applying a hash function to the extracted keyword; And constructing a hash function result value as a Bloom filter.

In one embodiment of the present invention, the plain text database may include numeric data and character data.

In one embodiment of the present invention, for a numerical data, various range type keywords can be generated.

In an embodiment of the present invention, the step of generating the encrypted tuple may generate the encrypted tuple by encrypting the tuple unit of the plain text database using a predetermined encryption key.

According to another aspect of the present invention, there is provided a method for searching a database encrypted by the above method in a database client terminal, the method comprising: extracting a keyword included in a plain- step; Applying a hash function to the extracted keyword; Modifying a query term so as to check whether a bit of the Bloom filter stored in the database server is set using the hash function result and transmitting the modified query term to the database server; Receiving, from the database server, a ciphertext corresponding to a Bloom filter in which a bit according to a hash function result included in the modified query term is set; And decrypting the received ciphertext.

According to another aspect of the present invention, there is provided a method of extracting statistical data from a database encrypted by the above method, comprising the steps of: extracting keywords included in a plain-text query input by a user; Applying a hash function to the extracted keyword; Modifying a query term so as to check whether a bit of the Bloom filter stored in the database server is set using the hash function result and transmitting the modified query term to the database server; And receiving the number of bloom filters having bits set according to a hash function result included in the modified query word from the database server.

The present invention as described above is effective in that the owner of the database possesses the encryption key and the decryption key to provide more secure encryption.

In addition, the present invention has an effect of applying a keyword to an element unit of a database to a hash function and constructing a bloom filter thereon to generate an index, thereby enabling real-time searching and statistics on the database.

1 is an exemplary view for explaining a Bloom filter.
2 is a conceptual diagram for explaining a database encryption method according to the present invention.
3 is a conceptual diagram for explaining a process of searching using an encryption database constructed according to the present invention.
FIG. 4 is a diagram for explaining a process of extracting statistical data using an encryption database constructed according to the present invention.
5 is a flowchart illustrating an encryption method according to an embodiment of the present invention.
6A to 6C are diagrams for explaining the results of the encryption method of the present invention.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

Hereinafter, a preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings.

The present invention proposes a method for improving a search function by using an index to which a Bloom filter is applied together with database encryption in units of tuples.

Further, in the present invention, when a database provided by an external cloud company is used without directly managing the database, the database manager of the external company intentionally uses the key managed by the database manager to store the sensitive information stored in the database We suggest a structure to make it impossible to browse. To this end, the invention allows the key owner to be owned only by the database owner, and the database administrator of the external vendor can only view the Bloom filter, which is used as an encryption database and index encrypted by the tuple, to decrypt the encrypted data I will not.

In addition, the present invention allows Bloom filter values in tuple units to be used for medical information analysis. It is possible to extract distribution data having a minimum error range by minimizing a false positive probability for the Bloom filter value. This makes it possible to know the statistics on the database information without browsing the encryption information. For example, the Bloom filter value in tuple units recorded in the database can be used to obtain the blood type distribution of men in their 50s living in Seoul. This protects personal information about medical information.

First, the Bloom filter will be described. A bloom filter is defined as a bit array that can check the presence or absence of data. 1 is an exemplary view for explaining a Bloom filter.

The type of data is divided into numeric data and character data. To retrieve character data, keywords that are data itself are used. On the other hand, in the case of numerical data, a numerical value indicating an interval is used. For example, in the case of a resident registration number, the top six bytes of the resident registration number means year / month / day. Therefore, it can be used as a 10-year unit keyword by using the upper first byte, and can be used as a year unit keyword by using the upper two bytes. In addition, the third and fourth bytes can be used as keywords for the month. Likewise, the first byte of the lower 7 bytes of the resident registration number can be used as a keyword indicating gender. Thus, by storing various numerical data in Bloom filter using various range type keywords, it can be useful for range calculation, aggregation calculation and statistical analysis.

First, the data required to search for data and the numerical values indicating various ranges are defined as data. In Figure 1, the data passes through k hash functions. The calculation result of the hash function is smaller than the size m of the Bloom filter, and the result of the hash function is stored in the Bloom filter.

The initial value of the Bloom filter is initialized to '0', and the corresponding bit is set to '1' according to the result of the hash function. The value of the stored Bloom filter is used to check for the presence of data. That is, in order to confirm the existence of data, it can be confirmed that the k-bit is set to '1' in the bloom filter according to the result of the hash function of the data. Also, if any one of k bits is set to '0', it can be known that no data exists.

However, since each bit of the bloom filter is set to only '0' and '1', it can be determined that there is also data that does not actually exist. This is called a false positive probability, and an appropriate size of m bits is used to minimize this probability.

Generally, when the size of the Bloom filter is m bits, k hash functions are used, and the number of data input to the Bloom filter is n, the positive error probability of the Bloom filter is expressed by the following equation.

The present invention uses the characteristics of the Bloom filter and uses the Bloom filter as an index for retrieving the encrypted data. That is, if it is determined that the data exists in the basic search using the Bloom filter, the decryption process is performed for the data encrypted in units of tuples, and a desired result can be found by removing the positive error part from the decrypted data.

Referring again to FIG. 1, FIG. 1 illustrates a method for building a Bloom filter.

First, all the numerical values representing the keywords and ranges required for data retrieval are defined as data, and these data pass through k hash functions. The hash function is a one-way function. The result of the calculation of the hash function is smaller than the size m of the bloom filter, and the result of the hash function is stored in the bloom filter.

For example, if the result of the hash function is 7, the seventh bit of the Bloom filter is set to one. Since the Bloom filter initially sets all the bits to 0, setting the Bloom filter to 1 according to the result of the hash function indicates that the corresponding data exists.

2 is a conceptual diagram for explaining a database encryption method according to the present invention.

As shown in the figure, the system to which the present invention is applied includes a DB client 20 corresponding to a database user part and a DB server 30. The DB client 20 is the owner side terminal of the data, and the DB server 30 may be a server of the cloud service provider responsible for data management. The DB client 20 and the DB server 30 may be connected through a network. The DB client 20 may be included in a user terminal (not shown).

The DB client 20 encrypts the contents of the plain text database 10 in units of tuples, calculates a hash function for each element in the tuple, and constructs a result of the hash function as a bloom filter.

 The encryption data and the Bloom filter in units of tuples are stored in the DB server 30 in the form of a table, respectively (see FIG. 6C). The encrypted data and the bloom field are stored as separate fields in the table, so that real-time retrieval is easy, and the user can more effectively cope with the join operation. In this case, the join operation is an operation for extracting related data from two or more tables. For example, it is an operation for searching for a student in 1970 in the personal information table and searching for a student having B or more in the grade table at the same time.

According to the present invention, since all the encryption processes and the Bloom filter configuration process are performed on a tuple-by-tuple basis, additional insertion or deletion is possible. The encrypted DB configuration part in the DB server 30 can be represented by an encrypted database E_tuple and a Bloom filter value bloom.

3 is a conceptual diagram for explaining a process of searching using an encryption database constructed according to the present invention.

The plain-text query created by the user 40 is transformed in the DB client 20 and transmitted to the DB server 30. [

The DB client 20 extracts keywords included in the plain-text query word, and calculates a hash function result value using the hash function. The plaintext query term is modified to check whether the bit of the Bloom filter stored in the DB server 30 is set as the result value of the calculated hash function. For example:

Plain text Query: Find someone with AB blood type living in Seoul.

Plain text query format:

  SELECT * FROM Std_info WHERE address = 'Seoul' and blood_type = 'AB'

Modified query format:

SELECT E_tuple FROM Std_info_bf

WHERE ((CONVERT (bit, SUBSTRING (bloom, [h1], 1)) & 1) = 1

AND (CONVERT (bit, SUBSTRING (bloom, [h2], 1) & 1) = 1

The DB server 30 finds the bloom filter in which the bits corresponding to the result values of the hash function for the keywords 'Seoul' and 'AB' included in the query entered from the user 40 are found and transmits the corresponding ciphertext to the DB client module 20).

The DB client 20 may decode the decrypted key using the decryption key to exclude the part where the positive error has occurred, and output the final result as a plain text.

FIG. 4 is a diagram for explaining a process of extracting statistical data using an encryption database constructed according to the present invention.

3, the DB client 20 modifies the plain text query input by the user 40 and transmits it to the DB server 30 as well.

That is, the DB client 20 extracts keywords included in the plain-text query, calculates a hash function result using the hash function, and outputs the result of the hash function as a bit of the bloom filter stored in the DB server 30 Is set, the plain-text query is modified.

The DB server 30 receives the modified query, finds the bloom filter in which bits corresponding to the hash function result value are set, confirms the number of bloom filters corresponding to the bloom filter, and transmits the bloom filter to the DB client 20.

At this time, unlike the procedure in FIG. 3, a decryption process for ciphertext data is not necessary, and statistical data can be extracted by the number of Bloom filters. At this time, the possible positive error probability can be minimized by using Equation (1). For example, by adjusting the c value using k = 1 and m = cn, the probability of positive error can be minimized. For c = 8, the probability of positive error can be reduced to less than 0.4%.

5 is a flowchart illustrating an encryption method according to an embodiment of the present invention.

As shown in the figure, in the database encryption method of the present invention, a keyword is extracted for each element unit of a plain text database (S1), and a hash function is applied to the extracted keyword (S2). The DB client 20 constructs a hash function result value using the hash function as a bloom filter (S3), and transmits the result value of the bloom filter to the DB server 30 (S4). S1 through S4 may be referred to as an index generation process for real-time searching of a database

The fact that character data is constituted by a Bloom filter using a keyword and that a numerical data is constituted by a Bloom filter using various range type keywords can be used for range calculation and aggregation calculation and statistical analysis, same.

On the other hand, the DB client 20 encrypts the tuple unit of the plaintext database using the encryption key (S5), and transmits the encrypted ciphertext unit encrypted to the DB server 30.

The DB server 30 can receive the ciphertext of the tuple unit and the result value of the bloom filter to construct an encryption database (S7).

6A to 6C are diagrams for explaining results according to the encryption method of the present invention. FIG. 6A shows plain text data, FIG. 6B shows data encrypted by an encryption method using a conventional bucket index, Represents data encrypted by the encryption method using the Bloom filter according to the present invention.

As shown in FIG. 6B, according to the encryption method using the conventional bucket index, it can be seen that buckets of the same size are used. Referring to FIG. 6C, the present invention is configured as a bloom filter using buckets of various sizes .

In the database as in the examples of FIGS. 6A to 6C, when the user inputs the query term 'Find students born in 1971', the results are shown in the following table.

Plain text Bucket index method Bloom filter method Processing time 0.033801 0.599700 0.264015 Number of results 3332 33363 3342

In the database shown in the example of FIGS. 6A to 6C, when the user inputs the query term 'Find students from 1980 to 1991,' the results are shown in the following table.

Plain text Bucket index method Bloom filter method Processing time 0.063836 1.190568 0.770244 Number of results 39972 66637 39987

As described above, according to the present invention, the processing time is much faster than the conventional bucket index method, and the number of retrieval results is also similar to the number of retrieval results for plaintext data.

Encrypting sensitive information is the best way to protect sensitive personal information stored in the database. However, once the database is encrypted, there is a significant barrier to real-time scanning. In addition, when the database is encrypted in the DB server, the encryption / decryption key may be intentionally leaked by the database administrator, and thus the encrypted content may be decrypted. Therefore, it is safe for the owner of the database to own the encryption and decryption keys.

In this way, the present invention is applied to a database managed by a database manager or a cloud service provider in a different environment, that is, a database managed by a subcontractor or the owner of the database, We suggest a possible method.

The present invention proposes a method for constructing an index using a Bloom filter for real-time retrieval of an encrypted database. A keyword is used for character data, and a Bloom filter is configured as an index using upper bytes for numeric data, It is proposed that the same value retrieval, range type retrieval, and aggregation operation are possible.

Further, in order to analyze the statistical data, the present invention can know the distribution by checking only the Bloom filter. For example, although it took a considerable time to process a range-type query such as 'blood type distribution of a girl born in the 1970's' in a general encrypted database, according to the present invention, a 10-year, 1-year, By using various kinds of keywords of the unit, since the keyword suitable for the range of the query term is stored in the Bloom filter, it is possible to perform real-time searching more quickly.

In addition, the present invention is capable of individually inserting and deleting data, and is easy to operate even when a database is managed by a subcontractor. Security can be effectively maintained even when a database is stored internally, This is possible.

While the invention has been shown and described with reference to certain preferred embodiments thereof, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined by the appended claims. Accordingly, the true scope of the present invention should be determined by the following claims.

20: DB client 30: DB server

Claims (10)

delete delete delete delete delete delete delete delete Comprising the steps of: constructing a Bloom filter for element units of a plaintext database including numeric data and character data to generate an index; Encrypting the tuple unit of the plain text database using a predetermined encryption key to generate the encrypted tuple; Transmitting the Bloom filter and the encrypted tuple to a database server; And storing the encrypted tuple and the Bloom filter as individual fields in a table, the database server constructing an encrypted database, the method comprising the steps of: The generating step may include: extracting a keyword for each element unit of the plain text database; Applying a hash function to the extracted keyword; And constructing a hash function result value as a Bloom filter,
Extracting a keyword included in a plain-text query input by a user;
Applying a hash function to the extracted keyword;
Transforming the query term so as to check whether the bit of the Bloom filter stored in the database server is set using the hash function result and transmitting the modified query term to the database server;
Receiving, from the database server, a ciphertext corresponding to a Bloom filter in which a bit according to a hash function result included in the modified query term is set; And
And decrypting the received ciphertext.
Comprising the steps of: constructing a Bloom filter for element units of a plaintext database including numeric data and character data to generate an index; Encrypting the tuple unit of the plain text database using a predetermined encryption key to generate the encrypted tuple; Transmitting the Bloom filter and the encrypted tuple to a database server; And storing the encrypted tuple and the Bloom filter as individual fields in a table, the database server constructing an encrypted database, the method comprising the steps of: The generating step may include: extracting a keyword for each element unit of the plain text database; Applying a hash function to the extracted keyword; And constructing a hash function result value as a Bloom filter,
Extracting a keyword included in a plain-text query input by a user;
Applying a hash function to the extracted keyword;
Transforming the query term so as to check whether the bit of the Bloom filter stored in the database server is set using the hash function result and transmitting the modified query term to the database server; And
And receiving a number of bloom filters in which a bit according to the hash function result included in the modified query term is set from the database server.

Images