SYSTEMS AND METHODS FOR A DISTRIBUTED QUERY EXECUTION ENGINE

• US 20160034529A1
• Filed: 07/23/2015
• Published: 02/04/2016
• Est. Priority Date: 07/29/2014
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Metanautix, Inc. (Microsoft Corporation)
• Inventors: Nguyen, Patrick An Phu, Kryze, David, Vassilakis, Theodore, Lerios, Apostolos
• Granted Patent: US 10,176,236 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24549 Run-time optimisation
• CPC Class Codes: G06F 16/2471 Distributed queries
• CPC Class Codes: G06F 16/33 Querying
• CPC Class Codes: G06F 16/3325 Reformulation based on results of preceding query
• CPC Class Codes: G06F 16/3338 Query expansion
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/36 Creation of semantic tools, e.g. ontology or thesauri
• CPC Class Codes: G06F 16/83 Querying
• CPC Class Codes: G06F 16/90335 Query processing

Abstract

Various embodiments of the present disclosure can include systems, methods, and non-transitory computer readable media configured to receive at least one database query to be executed. Code corresponding to the at least one database query can be generated. One or more optimizations to the generated code can be performed to produce specialized modular code. The one or more optimizations can include Just-In-Time (JIT) compilation techniques. Respective portions of the code can be distributed to a plurality of distributed computing systems for execution, wherein each of the distributed computing systems is connected to a portion of the plurality of distributed computing systems. A result for the at least one database query can be provided.

Claims

1. A computer-implemented method comprising:

receiving, by a computing system, at least one database query to be executed;

generating, by the computing system, code corresponding to the at least one database query;

performing, by the computing system, one or more optimizations to the generated code to produce specialized modular code, the one or more optimizations including Just-In-Time (JIT) compilation techniques;

distributing, by the computing system, respective portions of the code to a plurality of distributed computing systems for execution, wherein each of the distributed computing systems is connected to a portion of the plurality of distributed computing systems; and

providing, by the computing system, a result for the at least one database query.

2. The computer-implemented method of claim 1, wherein the code for the at least one database query is generated entirely on-the-fly.

3. The computer-implemented method of claim 1, wherein performing, by the computing system, one or more optimizations to the generated code further comprises:

performing, by the computing system, at least one of;

register allocation, inlining, constant folding, loop strength reduction, or loop-invariant code motion.

4. The computer-implemented method of claim 1, wherein performing, by the computing system, one or more optimizations to the generated code further comprises:

combining, by the computing system, the generated code with at least a portion of pre-compiled code.

5. The computer-implemented method of claim 4, wherein the code for the at least one database query is generated entirely on-the-fly and wherein the pre-compiled code is generated prior to receiving the at least one database query.

6. The computer-implemented method of claim 4, wherein the pre-compiled code is customized based at least in part on the code generated for the database query.

7. The computer-implemented method of claim 4, wherein the pre-compiled code is written in a high-level programming language.

8. The computer-implemented method of claim 4, wherein the pre-compiled code, when executed, performs vectorized instructions (SIMD) with a variable number of operands, and wherein combining, by the computing system, the generated code with at least a portion of pre-compiled code further comprises:

customizing, by the computing system, the vectorized instructions (SIMD) using one or more operands included in the database query.

9. The computer-implemented method of claim 1, wherein execution of the code is based at least in part on one or more iterator-based execution nodes, wherein respective code associated with the one or more iterator-based execution nodes is able to be suppressed upon determining that the respective code does not need to be executed.

10. The computer-implemented method of claim 9, wherein performing, by the computing system, the one or more optimizations to the generated code further comprises:

performing, by the computing system, one or more levels of code optimization, wherein each level of optimization introduces a further specialization to the code, and wherein each level of optimization is cached.

11. The computer-implemented method of claim 9, wherein at least one result associated with an iterator-based execution node resides in one or more processor registers and does not materialize into memory.

12. The computer-implemented method of claim 9, further comprising:

determining that the database query involves manipulation of one or more nested arrays; and

evaluating the one or more nested arrays based at least in part on the one or more iterator-based execution nodes.

13. The computer-implemented method of claim 12, wherein evaluating the one or more nested arrays further comprises:

processing each array in the one or more nested arrays as a series of rows.

14. The computer-implemented method of claim 1, wherein the computing system utilizes a caching mechanism that allows data items to be scheduled to the same distributed computing system across instances of a same query as the at least one database query.

15. The computer-implemented method of claim 1, wherein the computing system utilizes a caching mechanism that allows data items to be scheduled to the same distributed computing system across instances of different queries.

16. The computer-implemented method of claim 1, the method further comprising:

determining, by the computing system, that at least one of the distributed computing systems are non-responsive; and

re-distributing, by the computing system, respective portions of the optimized code corresponding to the at least one of the distributed computing system to one or more of the distributed computing systems that are responsive.

17. The computer-implemented method of claim 16, wherein determining, by the computing system, that at least one of the distributed computing systems are non-responsive further comprises:

determining, by the computing system, that at least one of the distributed computing systems are non-responsive while the at least one database query is being executed.

18. The computer-implemented method of claim 1, distributing, by the computing system, respective portions of the optimized code to a plurality of distributed computing systems for execution further comprises:

dispatching, by the computing system, the respective portions of the optimized code to a plurality of distributed computing systems based at least in part on a local chunked sort technique.

19. A system comprising:

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the system to perform;

receiving at least one database query to be executed;

generating code corresponding to the at least one database query;

performing one or more optimizations to the generated code;

distributing respective portions of the optimized code to a plurality of distributed computing systems for execution, wherein each of the distributed computing systems is connected to a portion of the plurality of distributed computing systems; and

providing a result for the at least one database query.

20. A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform:

receiving at least one database query to be executed;

generating code corresponding to the at least one database query;

performing one or more optimizations to the generated code;

distributing respective portions of the optimized code to a plurality of distributed computing systems for execution, wherein each of the distributed computing systems is connected to a portion of the plurality of distributed computing systems; and

providing a result for the at least one database query.