Using message-passing with procedural code in a database kernel

• US 9,684,685 B2
• Filed: 10/24/2013
• Issued: 06/20/2017
• Est. Priority Date: 10/24/2013
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP SE
• Inventors: Schreter, Ivan
• Primary Examiner(s): Saeed, Usmaan
• Assistant Examiner(s): Weinrich, Brian E.
• Application Number:
  US14/062,810
• Publication Number:
  US 20150120660A1
• Time in Patent Office: 1,335 Days
• Field of Search: 718105, 719314, 707620, 707626
• CPC Class Codes: G06F 16/21 Design, administration or maintenance of databases
• CPC Class Codes: G06F 16/23 Updating
• CPC Class Codes: G06F 16/2315 Optimistic concurrency control
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: G06F 16/275 Synchronous replication
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system
• CPC Class Codes: G06F 9/546 Message passing systems or structures, e.g. queues

Abstract

Each of a plurality of database objects can be assigned to a specific message-passing worker of a plurality of message-passing workers, each executing on a first logical core that shares with at least a second logical core one or more resources of a physical processor core of a plurality of physical processor cores. The second logical core can execute a job worker of a plurality of job workers that implement procedural code. Exclusive write access can be provided to a database object of the plurality of database objects via a message-passing worker of the plurality of message-passing workers while read-only access is provided to any database object of the plurality of database objects via any of the plurality of job workers. Operations can be executed by the message-passing worker in an order in which request messages are received in a message queue of the message-passing worker.

Claims

1. A computer program product comprising a non-transitory machine-readable medium storing instructions that, when executed by at least one programmable processor that comprises at least one physical core and a plurality of logical cores, cause the at least one programmable processor to perform operations comprising:

designating a first logical core from the plurality of logical cores for use by a first message-passing worker of a plurality of message-passing workers;

designating a second logical core from the plurality of logical cores for use by a first job worker of a plurality of job workers, the first and second logical cores sharing one or more resources of the at least one physical core;

designating an additional logical core from the plurality of logical cores for use by a second message-passing worker of the plurality of message-passing workers;

assigning a first set of database objects from a plurality of database objects in a database management system to the first message-passing worker such that the first message-passing worker has exclusive write access to the first set of database objects;

assigning a second set of database objects from the plurality of database objects in the database management system to the second message-passing worker such that the second message-passing worker has exclusive write access to the second set of database objects, the first set of database objects and the second set of database objects having at least one database object in common;

providing read-only access to the plurality of database objects via the first job worker;

executing operations by the first message-passing worker in an order in which request messages are received in a message queue of the first message-passing worker;

running a rebalancing operation to reassign the plurality of database objects to equalize a load across the plurality of message-passing workers;

receiving, by a rebalanced message-passing worker of the plurality of message-passing workers after the running of the rebalancing operation, a new request message pertaining to a reassigned database object of the plurality of database objects, the reassigned database object formerly being assigned to the rebalanced message-passing worker but currently being assigned to a new message-passing worker of the plurality of message-passing workers; and

posting, by the rebalanced message-passing worker, the new request message into a message queue of the new message-passing worker.

2. A computer program product as in claim 1, wherein the assigning comprises at least one of a round-robin assignment of the plurality of database objects to the plurality of message-passing workers and a random assignment of the plurality of database objects to the plurality of message-passing workers.

3. A computer program product as in claim 1, wherein the operations further comprise determining when to run the rebalancing operation, the determining comprising maintaining and monitoring performance statistics for each database object of the plurality of database objects.

4. A computer program product as in claim 1, wherein the read-only access is provided via a dirty read with no locks, and wherein data consistency are ensured using versioned objects and multi-version concurrency control to filter visible rows.

5. A system comprising:

at least one programmable processor comprising at least one physical core and a plurality of logical cores; and

a non-transitory machine-readable medium storing instructions that, when executed by the at least one programmable processor, cause the at least one programmable processor to perform operations comprising;

designating a first logical core from the plurality of logical cores for use by a first message-passing worker of a plurality of message-passing workers;

designating a second logical core from the plurality of logical cores for use by a first job worker of a plurality of job workers, the first and second logical cores sharing one or more resources of the at least one physical core;

designating an additional logical core from the plurality of logical cores for use by a second message-passing worker of the plurality of message-passing workers;

assigning a first set of database objects from a plurality of database objects in a database management system to the first message-passing worker such that the first message-passing worker has exclusive write access to the first set of database objects;

assigning a second set of database objects from the plurality of database objects in the database management system to the second message-passing worker such that the second message-passing worker has exclusive write access to the second set of database objects, the first set of database objects and the second set of database objects having at least one database object in common;

providing read-only access to the plurality of database objects via the first job worker;

executing operations by the first message-passing worker in an order in which request messages are received in a message queue of the first message-passing worker;

running a rebalancing operation to reassign the plurality of database objects to equalize a load across the plurality of message-passing workers;

receiving, by a rebalanced message-passing worker of the plurality of message-passing workers after the running of the rebalancing operation, a new request message pertaining to a reassigned database object of the plurality of database objects, the reassigned database object formerly being assigned to the rebalanced message-passing worker but currently being assigned to a new message-passing worker of the plurality of message-passing workers; and

posting, by the rebalanced message-passing worker, the new request message into a message queue of the new message-passing worker.

6. A system as in claim 5, wherein the assigning comprises at least one of a round-robin assignment of the plurality of database objects to the plurality of message-passing workers and a random assignment of the plurality of database objects to the plurality of message-passing workers.

7. A system as in claim 5, wherein the operations further comprise determining when to run the rebalancing operation, the determining comprising maintaining and monitoring performance statistics for each database object of the plurality of database objects.

8. A system as in claim 5, wherein the read-only access is provided via a dirty read with no locks, and wherein data consistency are ensured using versioned objects and multi-version concurrency control to filter visible rows.

9. A computer-implemented method comprising:

designating a first logical core from a plurality of logical cores of a data processor for use by a first message-passing worker of a plurality of message-passing workers;

designating a first logical core from the plurality of logical cores for use by a first message-passing worker of a plurality of message-passing workers;

designating a second logical core from the plurality of logical cores for use by a first job worker of a plurality of job workers, the first and second logical cores sharing one or more resources of the at least one physical core;

designating an additional logical core from the plurality of logical for use by a second message-passing worker of the plurality of message-passing workers;

assigning a first set of database objects from a plurality of database objects in a database management system to the first message-passing worker such that the first message-passing worker has exclusive write access to the first set of database objects;

assigning a second set of database objects from the plurality of database objects in the database management system to the second message-passing worker such that the second message-passing worker has exclusive write access to the second set of database objects, the first set of database objects and the second set of database objects having at least one database object in common;

providing read-only access to the plurality of database objects via the first job worker;

executing operations by the first message-passing worker in an order in which request messages are received in a message queue of the first message-passing worker;

running a rebalancing operation to reassign the plurality of database objects to equalize a load across the plurality of message-passing workers;

receiving, by a rebalanced message-passing worker of the plurality of message-passing workers after the running of the rebalancing operation, a new request message pertaining to a reassigned database object of the plurality of database objects, the reassigned database object formerly being assigned to the rebalanced message-passing worker but currently being assigned to a new message-passing worker of the plurality of message-passing workers; and

posting, by the rebalanced message-passing worker, the new request message into a message queue of the new message-passing worker.

10. A computer-implemented method as in claim 9, wherein the assigning comprises at least one of a round-robin assignment of the plurality of database objects to the plurality of message-passing workers and a random assignment of the plurality of database objects to the plurality of message-passing workers.

11. A computer-implemented method as in claim 9, wherein the operations further comprise determining when to run the rebalancing operation, the determining comprising maintaining and monitoring performance statistics for each database object of the plurality of database objects.

12. A computer-implemented method as in claim 9, wherein the read-only access is provided via a dirty read with no locks, and wherein data consistency are ensured using versioned objects and multi-version concurrency control to filter visible rows.

13. A computer-implemented method as in claim 9, wherein the assigning, the providing of exclusive write access to the database object, the providing of read-only access, and the executing are performed by at least one system comprising at least one programmable processor.