Analyzing and application or service latency

• US 7,921,410 B1
• Filed: 04/09/2007
• Issued: 04/05/2011
• Est. Priority Date: 04/09/2007
• Status: Active Grant

• Current Assignee: Micro Focus LLC (Open Text Corporation)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Wade, Gerald T., Symons, Julie A., Cohen, Ira, Green, John M.
• Primary Examiner(s): Zhen; Wei Y
• Assistant Examiner(s): UNG, LANNY N
• Application Number:
  US11/784,611
• Time in Patent Office: 1,457 Days
• Field of Search: 717/124
• US Class Current: 717/124
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 2201/87 Monitoring of transactions

Abstract

A method for analyzing a latency of a transaction performance is provided. The method includes receiving first transaction latency data which includes a transaction latency of a transaction and a first plurality of latency components that contribute to the transaction latency, receiving a definition of normality for the transaction latency, determining whether the transaction latency is normal or abnormal based at least on the definition of normality, upon the determining that the transaction latency is abnormal, determining whether there is a sufficient amount of the first transaction latency data based on a predefined criterion, upon the determining that the amount of the first transaction latency data is sufficient, computing a normal latency for each of the first plurality of latency components; and ranking the first plurality of latency components based on a degree of abnormality of each of the first plurality of latency components, which is based on the computed normal latency for the each latency component.

Claims

1. A method for analyzing a latency of a transaction performance, comprising:

receiving first transaction latency data which includes a transaction latency of a transaction and a first plurality of latency components that contribute to the transaction latency;

receiving a definition of normality for the transaction latency;

determining whether the transaction latency is normal or abnormal based at least on the definition of normality;

upon the determining that the transaction latency is abnormal, determining whether there is a sufficient amount of the first transaction latency data based on a predefined criterion;

upon the determining that the amount of the first transaction latency data is sufficient, computing a normal latency for each of the first plurality of latency components; and

ranking the first plurality of latency components based on a degree of abnormality of each of the first plurality of latency components, which is based on the computed normal latency for the each latency component.

2. The method of claim 1, further comprising:

receiving a policy for ranking the first plurality of latency components based on the degree of abnormality.

3. The method of claim 2, wherein ranking the first plurality of latency components comprises:

ranking the first plurality of latency components based on the policy for ranking.

4. The method of claim 1, wherein receiving the first transaction latency data comprises:

receiving a plurality of transaction latency data samples for a plurality of predefined time intervals, each sample includes therein the transaction latency and the first plurality of latency components monitored at each predefined time interval.

5. The method of claim 1, wherein receiving the definition of normality includes:

receiving a predefined threshold for determining a normality of the transaction latency data.

6. The method of claim 1, wherein receiving the policy for ranking the first plurality of latency components based on the degree of abnormality comprises:

receiving an instruction to rank each of the first plurality of latency components based on an amount of deviation of the each latency component from the computed normal latency of the each latency component.

7. The method of claim 1, wherein determining whether there is a sufficient amount of the transaction latency data comprises:

determining whether the first transaction latency data is sufficient for computing a normal latency for the at least one latency component therein as the predefined criterion.

8. The method of claim 4, wherein determining whether the transaction latency is normal or abnormal includes determining whether the transaction latency of each of the plurality of transaction latency data samples is normal or abnormal;

and the method further comprises;

upon the determining that the transaction latency of one of the plurality of transaction latency data samples is normal, adding the one transaction latency data sample to a training window for the determining of whether there is a sufficient amount of the first transaction latency data.

9. The method of claim 1, wherein computing the normal latency for each of the first plurality of latency components comprises:

computing the normal latency based at least on a mean value of a normal distribution for each of the first plurality of latency components.

10. The method of claim 1, further comprising:

determining how much contribution of each of the first plurality of latency components to the abnormal transaction latency of the second transaction latency data based on the ranking.

11. A method for analyzing a latency of a transaction performance, comprising:

receiving first transaction latency data which includes;

a) a transaction latency of a transaction;

b) a first indication that the transaction latency is normal or abnormal; and

c) a first plurality of latency components that contribute to the transaction latency;

determining whether the transaction latency is normal or abnormal based on the first indication in the first transaction latency data;

upon the determining that the transaction latency is abnormal, determining whether there is a sufficient amount of the transaction latency data in a training window;

upon the determining that the amount of the transaction latency data is sufficient in the training window, computing a normal latency for each of the first plurality of latency components based on the data in the training window; and

ranking the first plurality of latency components based on a degree of abnormality of each of the first plurality of latency components, which is based on the computed normal latency for the each latency component.

12. The method of claim 11, wherein receiving the first transaction latency data comprises:

receiving a first plurality of transaction latency data samples, each monitored at a predefined time interval and includes therein the transaction latency and the first plurality of latency components monitored at the predefined time interval.

13. The method of claim 12, wherein determining whether the transaction latency is normal comprises:

determining whether each of the first plurality of transaction latency data samples is normal based on the first indication in the each of the first plurality of transaction data samples.

14. The method of claim 13, wherein determining whether there is a sufficient amount of the first transaction latency data in a training window comprises:

determining whether there is a sufficient number of samples in the first plurality of transaction latency data samples that have therein the first indication of normal.

15. The method of claim 14, further comprising:

moving the training window by replacing a predetermined number of oldest data samples therein with a corresponding number of new data samples received subsequent to the receiving the first transaction latency data.

16. The method of claim 15, wherein computing a normal latency for each of the first plurality of latency components comprises:

computing a normal latency for each of the first plurality of latency components based on the data in the moving training window.

17. The method of claim 16, wherein the degree of abnormality of one of the second plurality of latency components is further based on a value of the latency component in reference to a computed normal latency of the one latency component as based on a latency-ranking policy.

18. The method of claim 17, wherein the first transaction latency data further includes at least one latency sub-component of one of the first plurality of latency components, the at least one latency sub-component contributes to the latency of both the one latency component and the transaction latency in the first transaction latency data.

19. A computer readable medium on which is encoded computer-executable programming code that includes computer execution instructions to:

receive first transaction latency data which includes a transaction latency of a transaction and a first plurality of latency components that contribute to the transaction latency;

receive a definition of normality for the transaction latency;

determine whether the transaction latency is normal or abnormal based at least on the definition of normality;

determine whether there is a sufficient amount of the first transaction latency data based on a predefined criterion upon the determining that the transaction latency is abnormal;

compute a normal latency for each of the first plurality of latency components upon the determining that the amount of the first transaction latency data is sufficient; and

rank the first plurality of latency components based on a degree of abnormality of each of the first plurality of latency components, which is based on the computed normal latency for the each latency component.

20. The computer-readable medium of claim 19, wherein the computer execution instructions to receive the first transaction latency data include:

computer-execution instructions to receive a plurality of transaction latency data samples for a plurality of predefined time intervals, each sample includes therein the transaction latency and the first plurality of latency components monitored at each predefined time interval.