Homomorphic database operations apparatuses, methods and systems

Patent metadata

Field	Value
Publication number	US-9912472-B2
Application number	US-201715628383-A
Country	US
Kind code	B2
Filing date	Jun 20, 2017
Priority date	Aug 1, 2013
Publication date	Mar 6, 2018
Grant date	Mar 6, 2018

Abstract

The HOMOMORPHIC DATABASE OPERATIONS APPARATUSES, METHODS AND SYSTEMS (“HEDO”) transform transaction storage requests and homomorphic model queries using HEDO components into homomorphic model query results. In some implementations, the disclosure provides a processor-implemented method of securely querying a shared homomorphically encrypted data repository and performing cross-table homomorphic joins.

First claim

What is claimed is:

1. An encrypted table value homomorphically joining apparatus, comprising:
a memory; and
a processor disposed in communication with said memory, and configured to issue a plurality of processing instructions stored in the memory, wherein the processor issues instructions to:
receive a model query for processing, wherein the model query requires data values from more than one table in a data repository;
based on the model query, determine a join type to perform, a plurality of tables to be joined, and a plurality of fields on which to join the plurality of tables;
determine that the plurality of fields on which to join includes at least one field that contains homomorphically encrypted data;
determine a homomorphic join strategy;
perform a homomorphic join using the plurality of fields; and
provide the resultant homomorphically joined tables,
wherein the homomorphic join strategy is directly comparing values in two homomorphically encrypted fields when it is determined that the fields have both been encrypted using a common deterministic encryption scheme.

2. The apparatus of claim 1 , wherein the processor further issues instructions to:
analyze the resultant homomorphically joined tables to determine that at least one record in the resultant homomorphically joined tables requires further processing; and
determine that the at least one record in the resultant homomorphically joined tables contains homomorphically encrypted data.

3. The apparatus of claim 2 , wherein the processor further issues instructions to perform a homomorphically optimized addition function to add the homomorphically encrypted data values contained in the at least one record.

4. The apparatus of claim 1 , wherein the homomorphic join strategy is using a key common to all values in at least one of the plurality of fields to decrypt the plurality of homomorphically encrypted values in a field.

5. The apparatus of claim 1 , wherein when the homomorphic join strategy is to utilize grouping keys, additionally comprising instructions to:
receive a plurality of grouping keys;
determine a plurality of ranges of field values that each grouping key may be used to decrypt; and
decrypt each of the field values in each of the ranges using the grouping keys.

6. The apparatus of claim 1 , additionally comprising instructions to aggregate results from the resultant homomorphically joined tables.

7. The apparatus of claim 6 , wherein aggregating involves leaving encrypted space and re-entering space.

8. The apparatus of claim 6 , wherein the aggregation is performed completely in encrypted space.

9. A processor implemented method of homomorphically joining repository tables using encrypted table values, comprising:
receiving, using one or more data processors, a model query for processing, wherein the model query requires data values from more than one table in a data repository;
based on the model query, determining, using the one or more data processors, a join type to perform, a plurality of tables to be joined, and a plurality of fields on which to join the plurality of tables;
determining, using the one or more data processors, that the plurality of fields on which to join includes at least one field that contains homomorphically encrypted data;
determining a homomorphic join strategy;
performing, using the one or more data processors, a homomorphic join using the plurality of fields; and
providing, using the one or more data processors, the resultant homomorphically joined tables,
wherein the homomorphic join strategy is directly comparing values in two homomorphically encrypted fields when it is determined that the fields have both been encrypted using a common deterministic encryption scheme.

10. The method of claim 9 , further comprising:
analyzing the resultant homomorphically joined tables to determine that at least one record in the resultant homomorphically joined tables requires further processing; and
determining that the at least one record in the resultant homomorphically joined tables contains homomorphically encrypted data.

11. The apparatus of claim 10 , wherein the processor further issues instructions to perform a homomorphically optimized addition function to add the homomorphically encrypted data values contained in the at least one record.

12. The method of claim 9 , wherein the homomorphic join strategy is using a key common to all values in at least one of the plurality of fields to decrypt the plurality of homomorphically encrypted values in a field.

13. The method of claim 9 , wherein when the homomorphic join strategy is to utilize grouping keys, additionally comprising:
receiving a plurality of grouping keys;
determining a plurality of ranges of field values that each grouping key may be used to decrypt; and
decrypting each of the field values in each of the ranges using the grouping keys.

14. The method of claim 9 , additionally comprising aggregating results from the resultant homomorphically joined tables.

15. The method of claim 14 , wherein aggregating involves leaving encrypted space and re-entering the encrypted space.

16. The method of claim 14 , wherein the aggregation is performed completely in encrypted space.

17. A non-transitory computer readable medium storing instructions that, when executed, cause an apparatus at least to perform:
receiving model query for processing, wherein the model query requires data values from more than one table in a data repository;
based on the model query a join type to perform, a plurality of tables to be joined, and a plurality of fields on which to join the plurality of tables;
determining that the plurality of fields on which to join includes at least one field that contains homomorphically encrypted data;
determining a homomorphic join strategy;
performing a homomorphic join using the plurality of fields; and
providing the resultant homomorphically joined tables,
wherein the homomorphic join strategy is directly comparing values in two homomorphically encrypted fields when it is determined that the fields have both been encrypted using a common deterministic encryption scheme.

18. The non-transitory computer readable medium of claim 17 further comprising instructions causing the apparatus to perform:
analyzing the resultant homomorphically joined tables to determine that at least one record in the resultant homomorphically joined tables requires further processing; and
determining that the at least one record in the resultant homomorphically joined tables contains homomorphically encrypted data.

19. The non-transitory computer readable medium of claim 18 further comprising instructions causing the apparatus to perform a homomorphically optimized addition function to add the homomorphically encrypted data values contained in the at least one record.

20. The non-transitory computer readable medium of claim 17 further comprising instructions causing the apparatus to perform aggregating results from the resultant homomorphically joined tables.

Assignees

• Visa Int Service Ass

Inventors

• Payton Paul
• Edington Scott
• Van Tilburg Johan

Classifications

• H04L9/008 Primary

  involving homomorphic encryption · CPC title

• H04L9/14

  using a plurality of keys or algorithms · CPC title