Method for identifying a workload type for a given workload of database requests

• US 7,499,908 B2
• Filed: 04/29/2003
• Issued: 03/03/2009
• Est. Priority Date: 04/23/2003
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Horman, Randall W., Van Boeschoten, Robin D., Elnaffar, Said, Martin, Pat, Schiefer, Bernhard K., Lightstone, Sam S.
• Primary Examiner(s): Pham; Khanh B
• Application Number:
  US10/425,171
• Publication Number:
  US 20040225631A1
• Time in Patent Office: 2,135 Days
• Field of Search: 707/104.1, 702/186, 705/10
• US Class Current: 1/1
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99936 Pattern matching access
• CPC Class Codes: Y10S 707/99942 Manipulating data structure, e.g. compression, compaction, compilation

Abstract

Workload type to be managed by a database management system (DBMS) is a key consideration in tuning the DBMS. Allocations for resources, such as main memory, can be very different depending on whether the workload type is Online Transaction Processing (OLTP) or Decision Support System (DSS). The DBMS also experiences changes in workload type during the normal processing cycle of the DBMS. It would be preferable for the database administrators to recognize the significant shifts of workload type that demand reconfiguring the DBMS to maintain acceptable levels of performance. Disclosed is a workload type classifier module, used by a DBMS, for recognizing workload types so that the DBMS may then manage or adjust its performance and reconfigure its resources accordingly. The classifier may be constructed based on the most significant workload characteristics that differentiate OLTP from DSS. The classifier is then used for identifying changes in workload types contained in a workload.

Claims

1. A method of identifying a workload type concentration for a given workload, the method comprising:

producing a plurality of performance snapshots of a training set, the training set comprising a plurality of training workloads, each training workload predetermined to elicit a particular database behavior associated with a known workload type, the workload type comprising one of an Online Transactional Processing (OLTP) type and a Decision Support System (DSS) type, an OLTP type characterized by short simple database queries with many concurrent users, a DSS type characterized by long complex database queries with few concurrent users, each performance snapshot comprising database behavior across a predetermined time interval and classified as one of OLTP and DSS;

extracting a set of attributes from the training set;

constructing a set of rules based on the classification of the plurality of performance snapshots and the set of attributes, the set of rules organized into a decision tree having a rule per tree node, each node of the decision tree configured for testing an attribute related to the workload type;

selecting a sample of the given workload; and

identifying the workload type concentration of the selected sample based on a comparison between a plurality of performance snapshots from the selected sample and the set of rules, the workload type concentration comprising a combination of OLTP and DSS, wherein a database may be more accurately tuned based on the workload type concentration, thereby enhancing database performance.

2. The method of claim 1 wherein the set of rules tests for attributes relates to the workload type.

3. The method of claim 1 further comprising determining whether the workload type concentration identification is acceptable.

4. The method of claim 1 further comprising identifying a change in the workload type.

5. The method of claim 1 further comprising validating the set of rules when drastic changes occur to the given workload.

6. The method of claim 1 further comprising using a data mining process to construct the set of rules.

7. A method to identify a workload type concentration for a given workload, the method comprising:

producing a plurality of performance snapshots of a training set, the training set comprising a plurality of training workloads, each training workload predetermined to elicit a particular database behavior associated with a known workload type, the workload type comprising one of an Online Transactional Processing (OLTP) type and a Decision Support System (DSS) type, an OLTP type characterized by short simple database queries with many concurrent users, a DSS type characterized by long complex database queries with few concurrent users, each performance snapshot comprising database behavior across a predetermined time interval and classified as one of OLTP and DSS;

extracting a set of attributes from the training set;

constructing a set of rules based on the classification of the plurality of performance snapshots and the set of attributes, the set of rules organized into a decision tree having a rule per tree node, each node of the decision tree for testing an attribute related to the workload type;

selecting a sample of the given workload;

producing a plurality of performance snapshots for the selected sample;

comparing the plurality of performance snapshots from the selected sample and the set of rules;

classifying each performance snapshot from the selected sample as one of OLTP and DSS; and

identifying the workload type concentration of the selected sample based on an OLTP to DSS ratio, the workload type concentration comprising a combination of OLTP and DSS, wherein a database may be more accurately tuned based on the workload type concentration, thereby enhancing database performance.

8. The method of claim 7 wherein the set of rules tests for attributes relates to the workload type.

9. The method of claim 7 further comprising determining whether the predicted workload type concentration identification is acceptable.

10. The method of claim 7 further comprising identifying a change in the workload type.

11. The method of claim 7 further comprising validating the set of rules when drastic changes occur to the given workload.

12. The method of claim 7 further comprising using a data mining process to construct the set of rules.