Pooling work across multiple transactions for reducing contention in operational analytics systems

Patent metadata

Field	Value
Publication number	US-9772911-B2
Application number	US-201514671639-A
Country	US
Kind code	B2
Filing date	Mar 27, 2015
Priority date	Mar 27, 2015
Publication date	Sep 26, 2017
Grant date	Sep 26, 2017

Abstract

A method includes scanning multiple incoming database transaction requests. Each transaction includes one or more operations. Operations are clustered into a set of combined operations based on type of operation constraints. Log records are prepared and written for re-performing operations upon system failures, and for undoing operations upon an operation or a transaction failing to be processed fully. Each set of combined operations are performed within a thread. Each update operation is marked for a transaction within which the update operation belongs. Recoverable update operations belonging to a plurality of transactions are performed within a single logical thread of execution.

First claim

What is claimed is:

1. A method comprising:
scanning a plurality of incoming transaction requests for a database, wherein each transaction comprises at least one operation;
clustering operations into at least one set of combined operations based on type of operation and database pages that the operations operate on;
preparing and writing log records for re-performing operations upon system failures, and for undoing operations upon an operation failing to be processed fully;
marking each update operation with an identifier of the transaction within which the update operation belongs; and
performing each set of combined operations within a logical thread of execution.

2. The method of claim 1 , wherein multiple operations operate on a single database page under a consolidated latch.

3. The method of claim 2 , further comprising:
for each database page that operations operate on, recording a value of a monotonically increasing counter; and
incrementing the counter when the log records are written,
wherein the counter value recorded on a particular database page indicates a subset of operations that have been operated on the particular database page.

4. The method of claim 1 , wherein transactions arise from separate nodes on a cluster of nodes.

5. The method of claim 4 , wherein the logical thread comprises an execution thread spread across multiple nodes or VM guests.

6. The method of claim 1 , wherein:
an operation comprises at least one of an insert, a delete, or a data value modification;
for an insert operation, operations are combined that are inserting rows of data into a same database table; and
for a delete operation, operations are combined that are deleting data from a same database table.

7. The method of claim 1 , wherein database queries use the transaction identifier marked in the update operation for determining whether to ignore an operation effect or not.

8. The method of claim 1 , wherein the logical thread comprises an execution thread or process on at least one of a single node or virtual machine (VM).

9. The method of claim 1 , wherein operations are combined that are executing a common prepared statement.

10. A computer program product for pooling work across multiple transactions, the computer program product comprising a computer readable storage device having program code embodied therewith, the program code executable by a processor to:
scan a plurality of incoming database transaction requests, wherein each transaction comprises at least one operation;
cluster operations into a set of combined operations based on type of operation constraints and database pages the operations operate on;
prepare and write log records for re-performing operations upon system failures, and for undoing operations upon an operation failing to be processed fully;
perform each set of combined operations within a thread;
mark each update operation for a transaction within which the update operation belongs; and
perform recoverable update operations belonging to a plurality of transactions within a single logical thread of execution.

11. The computer program product of claim 10 , wherein multiple operations operate on a single database page under a consolidated latch.

12. The computer program product of claim 11 , further comprising:
for each database page that operations operate on, recording a value of a monotonically increasing counter; and
incrementing the counter when the log records are written,
wherein the counter value recorded on a particular database page indicates a subset of operations that have been operated on the particular database page.

13. The computer program product of claim 10 , wherein transactions arise from separate nodes on a file system cluster.

14. The computer program product of claim 10 , wherein:
an operation comprises at least one of an insert, a delete, or a data value modification;
for an insert operation, operations are combined that are inserting rows of data into a same database table;
for a delete operation, operations are combined that are deleting data from a same database table; and
database queries use the marked update operation for a transaction for determining whether to ignore an operation effect or not.

15. The computer program product of claim 14 , wherein the logical thread comprises an execution thread or process on at least one of a single node or virtual machine (VM) and the logical thread comprises an execution thread spread across multiple nodes or VM guests.

16. A system comprising:
a storage device configured to store one or more databases;
a processor configured to:
scan a plurality of incoming transaction requests for a database, wherein each transaction comprises at least one operation;
cluster operations into a set of combined operations based on type of operation constraints and database pages that the operations operate on, and prepares and writes log records for re-performing operations upon system failures, and for undoing operations upon an operation failing to be processed fully; and
marks each update operation for a transaction within which the update operation belongs;

a batched requests processor that performs each set of combined operations within a thread; and
a logger and roll-back processor that performs recoverable update operations belonging to a plurality of transactions within a logical thread of execution.

17. The system of claim 16 , wherein:
multiple operations operate on a single database page under a consolidated latch;
for each database page that operations operate on, the processor records a value of a monotonically increasing counter, and increments the counter when the log records are written; and
the counter value recorded on a particular database page indicates a subset of operations that have been operated on the particular database page.

18. The system of claim 16 , wherein:
transactions arise from separate nodes on a file system cluster;
an operation comprises at least one of an insert, a delete, or a data value modification;
for an insert operation, the cluster processor combines operations that are inserting rows of data into a same database table; and
for a delete operation, the cluster processor combines operations that are deleting data from a same database table,

wherein database queries use the marked update operation for a transaction for determining whether to ignore an operation effect or not.

19. The system of claim 18 , wherein the logical thread comprises an execution thread or process on at least one of a single node or virtual machine (VM).

20. The system of claim 18 , wherein the logical thread comprises an execution thread spread across multiple nodes or VM guests.

Assignees

• IBM

Inventors

• Barber Ronald J
• Bhattacharjee Bishwaranjan
• Hamedani Mohammad Sadoghi
• Lohman Guy M
• Mohan Chandrasekaran
• Pandis Ippokratis
• Raman Vijayshankar
• Sidle Richard S
• Storm Adam J

Classifications

• G06F9/466

  Transaction processing · CPC title

• G06F2201/87

  Monitoring of transactions · CPC title

• G06F11/1471

  involving logging of persistent data for recovery · CPC title

• G06F2201/80

  Database-specific techniques · CPC title

• G06F11/1446

  Point-in-time backing up or restoration of persistent data · CPC title