Performance degradation root cause prediction in a distributed computing system

• US 20070271219A1
• Filed: 05/17/2006
• Published: 11/22/2007
• Est. Priority Date: 05/17/2006
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Neogi, Anindya, Sailer, Anca, Agarwal, Manoj K., Kar, Gautam
• Granted Patent: US 7,412,448 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/008 Reliability or availability analysis

Abstract

A method of identifying at least one resource in a distributed computing system which is a potential root cause of performance degradation of the system includes the steps of: computing operational bounds for the system and operational bounds for at least one resource in the system; comparing current end-to-end system performance with the operational bounds for the system; when the current end-to-end system performance is outside of the operational bounds for the system, comparing current performance of the at least one resource in the system with the operational bounds for the at least one resource; and generating at least one output identifying the at least one resource in the system which is a potential root cause of performance degradation of the system.

Claims

1. A method of identifying at least one resource in a distributed computing system which is a potential root cause of performance degradation of the system, the method comprising the steps of:

computing operational bounds for the system and operational bounds for at least one resource in the system;

comparing current end-to-end system performance with the operational bounds for the system;

when the current end-to-end system performance is outside of the operational bounds for the system, comparing a current performance of the at least one resource in the system with the operational bounds for the at least one resource; and

generating at least one output identifying the at least one resource in the system which is a potential root cause of performance degradation of the system.

2. The method of claim 1, wherein at least one of the operational bounds for the system and the operational bounds for the at least one resource are calculated based on load-adjusted response time.

3. The method of claim 1, further comprising the step of computing an individual contribution of the at least one resource to an end-to-end response time for the system.

4. The method of claim 3, wherein the step of computing the individual contribution of the at least one resource to an end-to-end response time for the system comprises separately computing an individual contribution of the at least one resource for successful transactions and an individual contribution of the at least one resource for failed transactions.

5. The method of claim 3, wherein the step of computing an individual contribution of the at least one resource to an end-to-end response time for the system comprises using a dependency tree to determine which resources to inspect for a given transaction.

6. The method of claim 1, wherein the output is an ordered set of the at least one resource which is a potential root cause of performance degradation in the system.

7. The method of claim 1, wherein the output is sorted based at least in part on a comparison of the current performance of the at least one resource with the operational bounds of the at least one resource.

8. The method of claim 1, wherein the output is sorted based at least in part on a ratio of the current performance of the at least one resource and the operational bounds of the at least one resource.

9. Apparatus for identifying at least one resource in a distributed computing system which is a potential root cause of performance degradation of the system, the apparatus comprising:

a memory; and

at least one processor coupled to the memory and operative;

(i) to compute operational bounds for the system and operational bounds for at least one resource in the system;

(ii) to compare current end-to-end system performance with the operational bounds for the system;

(iii) when the current end-to-end system performance is outside of the operational bounds for the system, to compare current performance of the at least one resource in the system with the operational bounds for the at least one resource; and

(iv) to generate at least one output identifying the at least one resource in the system which is a potential root cause of performance degradation of the system.

10. The apparatus of claim 9, wherein at least one of the operational bounds for the system and the operational bounds for the at least one resource are calculated based on load-adjusted response time.

11. The apparatus of claim 9, wherein the processor is further operative to compute an individual contribution of the at least one resource to an end-to-end response time for the system.

12. The apparatus of claim 11, wherein the step of computing the individual contribution of the at least one resource to an end-to-end response time for the system comprises separately computing an individual contribution of the at least one resource for successful transactions and an individual contribution of the at least one resource for failed transactions.

13. The apparatus of claim 11, wherein the step of computing an individual contribution of the at least one resource to an end-to-end response time for the system comprises using a dependency tree to determine which resources to inspect for a given transaction.

14. The method of claim 9, wherein the output is an ordered set of the at least one resource which is a potential root cause of performance degradation in the system.

15. The apparatus of claim 9, wherein the output is sorted based at least in part on a ratio of the current performance of the at least one resource and the operational bounds of the at least one resource.

16. The apparatus of claim 9, wherein the output is sorted based at least in part on a comparison of the current performance of the at least one resource with the operational bounds of the at least one resource.

17. An article of manufacture for identifying at least one resource in a distributed system which is a potential root cause of performance degradation of the system, comprising a machine readable medium containing one or more programs which when executed implement the steps of:

computing operational bounds for the system and operational bounds for at least one resource in the system;

comparing current end-to-end system performance with the operational bounds for the system;

when the current end-to-end system performance is outside of the operational bounds for the system, comparing a current performance of the at least one resource in the system with the operational bounds for the at least one resource; and

generating at least one output identifying the at least one resource in the system which is a potential root cause of performance degradation of the system.