System and methods for processing data queries in hybrid data mesh architectures

Patent metadata

Field	Value
Publication number	US-12393590-B2
Application number	US-202218046858-A
Country	US
Kind code	B2
Filing date	Oct 14, 2022
Priority date	Oct 14, 2022
Publication date	Aug 19, 2025
Grant date	Aug 19, 2025

Abstract

System and methods for processing data queries in hybrid data mesh architectures are disclosed. A method for processing data queries in hybrid data mesh architectures may include an in-memory query engine: (1) receiving, from a requestor, a data query to retrieve data; (2) parsing the data query into a parse tree; (3) converting the parse tree into a relational tree, wherein the relational tree is a tree representation of relational operators used to execute the data query; and (4) executing a data retrieval method selected from the group consisting of tree partitioning by cost and pruning using subtree replacement using a plurality of retrieval services to retrieve data from data stores identified by the data retrieval method.

First claim

What is claimed is:

1. A method for processing data queries in hybrid data mesh architectures, comprising:
receiving, by an in-memory query engine and from a requestor, a data query to retrieve data and an identifier for a data location, wherein the data query further comprises a data retrieval budget, and an amount of compute resources used in the data retrieval is based on the data retrieval budget;
parsing, by the in-memory query engine, the data query into a parse tree;
converting, by the in-memory query engine, the parse tree into a relational tree, wherein the relational tree is a tree representation of relational operators used to execute the data query;
identifying, by the in-memory query engine, a data retrieval method selected from the group consisting of tree partitioning by cost and pruning using subtree replacement;
retrieving, by the in-memory query engine, data from the data location using a plurality of retrieval services and the data retrieval method; and
enforcing data restrictions associated with the identifier during the retrieval.

2. The method of claim 1 , further comprising:
aggregating, by the in-memory query engine, data retrieved by the plurality of retrieval services; and
returning, by the in-memory query engine, the aggregated data to the requestor.

3. The method of claim 1 , wherein the identifier comprises a Uniform Resource Identifier (URI), an Internet Protocol (IP) Address, and/or an executable location.

4. The method of claim 1 , wherein the in-memory query engine converts the parse tree into an abstract syntax tree (AST) and converts the AST into the relational tree.

5. The method of claim 1 , further comprising:
validating, by the in-memory query engine, a syntax of the data query using an in-memory representation of a plurality of data stores for the data location.

6. The method of claim 1 , wherein the in-memory query engine executes the tree partitioning by cost by iterating through all legal relational algebraic combinations of the relational tree and selecting a subtree from the iterations that has a lowest cost using a planner process.

7. The method of claim 1 , wherein the in-memory query engine executes pruning using subtree replacement by matching specific relational subtrees with a single relational node using a tree traversal.

8. The method of claim 1 , wherein the in-memory query engine and/or the plurality of retrieval services are executed in a secure enclave.

9. A method for schema discovery in hybrid data mesh architectures, comprising:
establishing, by an in-memory query engine, a connection with a plurality of data stores;
querying, by the in-memory query engine and by using a plurality of data retrieval services, each of the plurality of data stores for schema information for tables that are stored in the data store;
receiving, by the in-memory query engine, metadata comprising the schema information from the plurality of data stores; and
building, by the in-memory query engine, an in-memory representation of each table and a map representing a location of each table using the metadata.

10. The method of claim 9 , wherein the plurality of data stores are identified in a catalog that identifies each of the plurality of data stores by name and location.

11. The method of claim 9 , wherein the schema information comprises table name for tables, column names in the tables, and column data types for each of the columns.

12. The method of claim 9 , wherein the in-memory query engine and/or the plurality of retrieval services are executed in a secure enclave.

13. A non-transitory computer readable storage medium, including instructions stored thereon, which when read and executed by one or more computer processors, cause the one or more computer processors to perform steps comprising:
receiving, from a requestor, a data query retrieve data and an identifier for a data location, wherein the data query further comprises a data retrieval budget, and an amount of compute resources used in the data retrieval is based on the data retrieval budget;
parsing the data query into a parse tree;
converting the parse tree into a relational tree, wherein the relational tree is a tree representation of relational operators used to execute the data query;
identifying a data retrieval method selected from the group consisting of tree partitioning by cost and pruning using subtree replacement;
retrieving data from the data location using a plurality of data retrieval services and the data retrieval method;
enforcing data restrictions associated with the identifier during the retrieval;
aggregating data retrieved by the plurality of retrieval services; and
returning the aggregated data to the requestor.

14. The non-transitory computer readable storage medium of claim 13 , wherein the identifier comprises a Uniform Resource Identifier (URI), an Internet Protocol (IP) Address, and/or an executable location.

15. The non-transitory computer readable storage medium of claim 14 , further including instructions stored thereon, which when read and executed by one or more computer processors, cause the one or more computer processors to aggregate data retrieved by the plurality of retrieval services and return the aggregated data to the requestor.

16. The non-transitory computer readable storage medium of claim 13 , further including instructions stored thereon, which when read and executed by one or more computer processors, cause the one or more computer processors to execute the tree partitioning by cost by iterating through all legal relational algebraic combinations of the relational tree and selecting a subtree from the iterations that has a lowest cost using a planner process.

17. The non-transitory computer readable storage medium of claim 13 , further including instructions stored thereon, which when read and executed by one or more computer processors, cause the one or more computer processors to execute pruning using subtree replacement by matching specific relational subtrees with a single relational node using a tree traversal.

Assignees

• Jpmorgan Chase Bank Na

Inventors

• Tummala Narayana R
• Pistoia Marco
• Ligman Joseph W
• Azanli Elvir
• Deshpande Pranav

Classifications

• G06F16/2246

  Trees, e.g. B+trees · CPC title

• G06F16/258

  Data format conversion from or to a database · CPC title

• G06F16/24526

  Internal representations for queries · CPC title

• G06F16/24556 Primary

  Aggregation; Duplicate elimination · CPC title

• G06F16/2471 Primary

  Distributed queries · CPC title