CLOUD SERVICE INTERDEPENDENCY RELATIONSHIP DETECTION

• US 20180287877A1
• Filed: 03/31/2017
• Published: 10/04/2018
• Est. Priority Date: 03/31/2017
• Status: Active Grant

• Current Assignee: BMC Software Incorporated (KKR & Co., Inc.)
• Original Assignee: BMC Software Incorporated (KKR & Co., Inc.)
• Inventors: NAVARE, Nitin, SODEM, Sridhar, SCARPELLI, Joseph, RIVINGTON, Mark
• Granted Patent: US 10,931,532 B2
• Time in Patent Office: Days
• CPC Class Codes: H04L 41/12 Discovery or management of network topologies
• CPC Class Codes: H04L 41/142 using statistical or mathematical methods
• CPC Class Codes: H04L 41/5058 Service discovery by the service manager
• CPC Class Codes: H04L 43/0817 by checking functioning
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30

Abstract

A computer system includes a processor, a memory, a data collector, a relationships analyzer, and a topological map generator. The data collector retrieves performance data in a specific set of performance categories for computing resources in a computing system for a time interval. The relationships analyzer, for each computing resource-to-computing resource pair in the computing system, performs a correlation analysis of the respective behavior values of the computing resources in the pair, and identifies the computing resource-to-computing resource pairs that have correlation values exceeding a pre-determined threshold level as having performance interdependencies. The topological map generator prepares an undirected graph of the computing resources that have performance interdependencies, and displays the undirected graph as a topographic map of the computing resources in the computing system.

Claims

1. A method comprising:

retrieving performance data of computing resources in a computing system for a time interval, the performance data including performance data in one or more performance categories;

for each computing resource-to-computing resource pair in the computing system, performing a correlation analysis of the respective behavior values of the computing resources in the pair;

identifying the computing resource-to-computing resource pairs that have correlation values exceeding a pre-determined threshold level as having performance interdependencies;

preparing an undirected graph of the computing resources that have performance interdependencies; and

displaying, on a computer display, the undirected graph of the computing resources that have performance interdependencies as a topographic map of the computing resources in the computing system.

2. The method of claim 1 further comprising:

after performing the correlation analysis, performing a cross-correlation analysis of the respective behavior values to determine a time lag value that corresponds to a maximum correlation value.

3. The method of claim 2 further comprising:

based on the cross-correlation analysis, excluding computing resource-to-computing resource pairs that are cross-correlated with a non-zero lag and retaining only computing resource-to-computing resource pairs that are cross-correlated with a time lag of about zero for further processing.

4. The method of claim 1, wherein preparing the undirected graph of the computing

resources that have performance interdependencies includes:

preparing an adjacency matrix representing a finite graph from data identifying the computing resource-to-computing resource pairs that have correlation values exceeding the pre-determined threshold level as having performance interdependencies.

5. The method of claim 4, wherein the pre-determined threshold level corresponds to a correlation value of about zero.

6. The method of claim 1, wherein the one or more performance categories include one or more of central processing unit (CPU) utilization, disk reads, disk writes, memory utilization, load average, processor queue length, context switches, and network bandwidth usage.

7. The method of claim 1 further comprising:

before performing the correlation analysis, performing dimensionality reduction of the retrieved performance data using principal component analysis to arrive at a single behavior value for each of the computing resources in the computing system.

8. The method of claim 1 further comprising:

updating the undirected graph to represent a current status of the computing resources in the computing system.

9. A computer system comprising:

a processor;

a memory;

a data collector configured to;

retrieve performance data of computing resources in a computing system for a time interval, the performance data including performance data in one or more performance categories;

a relationships analyzer configured to;

for each computing resource-to-computing resource pair in the computing system, perform a correlation analysis of the respective behavior values of the computing resources in the pair; and

identify the computing resource-to-computing resource pairs that have correlation values exceeding a pre-determined threshold level as having performance interdependencies; and

a topological map generator configured to;

prepare an undirected graph of the computing resources that have performance interdependencies; and

display, on a computer display, the undirected graph of the computing resources that have performance interdependencies as a topographic map of the computing resources in the computing system.

10. The computer system of claim 9, wherein the relationships analyzer is configured to:

after performing the correlation analysis, perform a cross-correlation analysis of the respective behavior values to determine a time lag value that corresponds to a maximum correlation value.

11. The computer system of claim 10, wherein the relationships analyzer is further configured to:

based on the cross-correlation analysis, exclude computing resource-to-computing resource pairs that are cross-correlated with a non-zero lag and retain only computing resource-to-computing resource pairs that are cross-correlated with a time lag of about zero for further processing.

12. The computer system of claim 9, wherein the topological map generator is configured to:

prepare an adjacency matrix representing a finite graph from data identifying the computing resource-to-computing resource pairs that have correlation values exceeding the pre-determined threshold level as having performance interdependencies.

13. The computer system of claim 12, wherein the pre-determined threshold level corresponds to a correlation value of about zero.

14. The computer system of claim 9, wherein the one or more performance categories include one or more of CPU utilization, disk reads, disk writes, memory utilization, load average, processor queue length, context switches, and network bandwidth usage.

15. The computer system of claim 9, wherein the relationships analyzer is configured to:

before performing the correlation analysis, perform dimensionality reduction of the retrieved performance data using principal component analysis to arrive at a single behavior value for each of the computing resources in the computing system.

16. The computer system of claim 9, wherein the topological map generator is further configured to update the undirected graph to represent a current status of the computing resources in the computing system.

17. A computer program product including instructions recorded on a non-transitory computer-readable storage medium and configured to cause at least one processor to:

retrieve performance data of computing resources in a computing system for a time interval, the performance data including performance data in one or more performance categories;

for each computing resource-to-computing resource pair in the computing system, perform a correlation analysis of the respective behavior values of the computing resources in the pair;

identify the computing resource-to-computing resource pairs that have correlation values exceeding a pre-determined threshold level as having performance interdependencies;

prepare an undirected graph of the computing resources that have performance interdependencies; and

display, on a computer display, the undirected graph of the computing resources that have performance interdependencies as a topographic map of the computing resources in the computing system.

18. The computer program product of claim 17, wherein the instructions, when executed, are further configured to cause the at least one processor to:

after performing a correlation analysis, perform a cross-correlation analysis of the respective behavior values to determine a time lag value that corresponds to a maximum correlation value.

19. The computer program product of claim 18, wherein the instructions, when executed, are further configured to cause the at least one processor to:

based on the cross-correlation analysis, exclude computing resource-to-computing resource pairs that are cross-correlated with a non-zero lag and retain only computing resource-to-computing resource pairs that are cross-correlated with a time lag of about zero for further processing.

20. The computer program product of claim 17, wherein the instructions, when executed, are further configured to cause the at least one processor to:

when preparing the undirected graph of the computing resources that have performance interdependencies, prepare an adjacency matrix representing a finite graph from data identifying the computing resource-to-computing resource pairs that have correlation values exceeding the pre-determined threshold level as having performance interdependencies.

21. The computer program product of claim 20, wherein the pre-determined threshold level corresponds to a correlation value of about zero.

22. The computer program product of claim 17, wherein the specific set of performance categories includes one or more of CPU utilization, disk reads, disk writes, memory utilization, load average, processor queue length, context switches, and network bandwidth usage.

23. The computer program product of claim 17, wherein the instructions, when executed, are further configured to cause the at least one processor to:

before performing the correlation analysis, perform dimensionality reduction of the retrieved performance data using principal component analysis to arrive at a single behavior value for each of the computing resources in the computing system.

24. The computer program product of claim 17, wherein the instructions, when executed, are further configured to cause the at least one processor to:

update the undirected graph to represent a current status of the computing resources in the computing system.