Method for first pass filtering of anomalies and providing a base confidence level for resource usage prediction in a utility computing environment

• US 7,409,314 B2
• Filed: 07/29/2004
• Issued: 08/05/2008
• Est. Priority Date: 07/29/2004
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Childress, Rhonda L., Gasca, Miguel E. Jr., Tittizer, Abigail A., Pagan, Elfred
• Primary Examiner(s): Barlow, Jr.; John E
• Assistant Examiner(s): TEIXEIRA MOFFAT, JONATHAN CHARLES
• Application Number:
  US10/902,591
• Publication Number:
  US 20060025950A1
• Time in Patent Office: 1,468 Days
• Field of Search: None
• US Class Current: 702/182
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 2201/81 Threshold
• CPC Class Codes: G06F 2201/86 Event-based monitoring

Abstract

Provided is a method for monitoring resources in a utility computing environment (UCE). Measurements are evaluated to determine whether or not a particular resource requires remedial or other type of action. A sliding measurement window is employed to assemble a number of measurements corresponding to a particular resource. The number of intervals in a sliding measurement window is based upon best practices corresponding to the resource being measured and analyzed. A first threshold-crossing event and subsequent events are stored until the window is full, or closed. When the window is closed, the threshold-crossing measurements are analyzed to determine whether or not there exists an issue with the resource that requires action. Once a window has been closed and analyzed, the first threshold-crossing event and each subsequent event up to a second threshold-crossing event are discarded and the window reopens.

Claims

1. A method of detecting resource events in a utility computing environment, the method comprising:

periodically retrieving a measurement corresponding to a resource;

defining a first threshold for the resource measurement;

defining a duration attribute;

detecting a first measurement that crosses the threshold;

storing in a memory the first measurement and an indicator of threshold crossing;

for a number of periods corresponding to a second threshold, periodically retrieving subsequent measurements corresponding to the resource and storing in the memory the measurements and corresponding indicators of whether or not the subsequent measurements cross the first threshold; and

signaling a resource event if the number of stored threshold crossing measurements, which are part of any contiguous group of threshold crossing events such that the number of threshold crossing events in the corresponding contiguous group is greater than the duration attribute, exceed a third threshold after said number of periods.

2. The method of claim 1, further comprising investigating the functionality of the resource if a resource event is signaled.

3. The method of claim 1, further comprising generating a report on the first and subsequent measurements and the resource if a resource event is signaled.

4. The method of claim 1, further comprising,

discarding the first measurement and corresponding indicator and each subsequent measurement and corresponding indicator preceding a second measurement in which the threshold was crossed;

and

continuing the retrieving, storing and signaling.

5. The method of claim 1, further comprising,

determining, after the number of periods, whether or not only one threshold crossing measurement is stored;

discarding all the measurements and corresponding indicators if only one threshold crossing measurement is stored;

continuing the periodically retrieving and detecting; and

continuing the storing of measurements and indicators once a measurement crosses the first threshold.

6. The method of claim 1, wherein a score is calculated based upon the duration attributes.