ROOT-CAUSE APPROACH TO PROBLEM DIAGNOSIS IN DATA NETWORKS

• US 20090028053A1
• Filed: 07/27/2007
• Published: 01/29/2009
• Est. Priority Date: 07/27/2007
• Status: Active Grant

• Current Assignee: eG Innovations Pte Ltd.
• Original Assignee: eG Innovations Pte Ltd.
• Inventors: Kannan, Raja, Subramanian, Sreedharan, Vaidhinathan, Balamurugan, Ramanathan, Srinivas
• Granted Patent: US 8,208,381 B2
• Time in Patent Office: Days
• US Class Current: 370/241
• CPC Class Codes: G06F 11/0712 in a virtual computing platform, e.g. logically partitioned systems
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36

Abstract

An improved root-cause approach to problem diagnosis in data networks in the form of a method comprising the steps of: associating each metric in a at least one set of metrics with at least one component and/or network device; obtaining values for each such metric from a monitoring system; determining whether each such metric is indicative of a problem within the data network; and ranking and correlating indicative problems to determine whether a problem may be symptomatic of another problem based on an interconnection and/or interdependency between a physical machine and a virtual machine, between components or between components and network devices.

Claims

1. An improved root-cause approach to problem diagnosis in data networks in the form of a method comprising the steps of:

associating each metric in a at least one set of metrics with at least one component and/or network device

obtaining values for each such metric from a monitoring system;

determining whether each such metric is indicative of a problem within the data network; and

ranking and correlating indicative problems to determine whether a problem may be symptomatic of another problem based on an interconnection and/or interdependency between a physical machine and a virtual machine, between components or between components and network devices.

2. A method according to claim 1, further including the step of obtaining and storing information on the interconnection and interdependencies between each physical machine and virtual machine in the data network and where the step of ranking and correlating indicative problems includes the sub-step of ranking and correlating indicative problems to determine whether a problem may be symptomatic of another problem based on the stored physical/virtual machine interconnection and interdependency information.

3. A method according to claim 1 or claim 2, further including the step of obtaining and storing information on the interconnection and interdependencies between components and between components and network devices and where the step of ranking and correlating indicative problems includes the sub-step of performing end-to-end correlation of the indicative problems based on the stored component/network interconnection and interdependency information.

4. A method according to any preceding claim including the steps of:

defining a hierarchical layer model;

associating each layer in the hierarchical layer model with a set of metrics,

where the step of ranking and correlating indicative problems includes the sub-step of performing top-to-bottom correlation of the indicative problems based, at least in part, on the hierarchical layer model.

5. A method according to any one of claims 2 to 4, as dependent on claim 2, where the step of obtaining information on the interconnection and interdependencies between each physical machine and virtual machine in the data network is repeatedly obtained and compared to the prior stored information on the interconnection and interdependencies between each physical machine and virtual machine in the data network and if the information so obtained differs from the prior stored information, storing the obtained information in place of the prior stored information.

6. A method according to any one of claims 3 to 5, as dependent on claim 3, where the step of obtaining information on the interconnection and interdependencies between components and between components and network devices is repeatedly obtained and compared to the prior stored information on the interconnection and independencies between components and between components and network devices and if the information so obtained differs from the prior stored information, storing the obtained information in place of the stored information.

7. A method according to any preceding claim, including the step of defining a set of priority queues and the step of ranking the indicative problems is made with reference to the priority queues.

8. A method according to claim 7, as dependent on claim 3, where the step of performing end-to-end correlation is performed on each priority queue in turn.

9. A method according to claim 7 or claim 8, as dependent on claim 4, where the step of performing top-to-bottom correlation is performed between each priority queue and the next lowest priority queue, if any.

10. A method according to any one of claims 7 to 9, including the step of demoting any indicative problem determined to be symptomatic of another indicative problem to the next lowest priority queue where possible to do so.

11. A method according to any one of claims 4 to 10, where those indicative problems relating to the layer upon which all other layers are dependent are processed first.

12. A method according to any preceding claim, where each indicative problem is associated with a component name which includes a port number and each metric is associated with a port number, the method including the step of comparing the port number of the component name with the port number of the associated metric and, if the comparison shows that the indicative problem is independent of the port number, the component name is modified to delete the associated port number.

15. A computer readable medium in accordance with claim 14, further operable to perform the method as claimed in claims 2 to 12.

13. A system for implementing an improved root-cause approach to problem diagnosis in data networks comprising:

a monitoring system; and

an analytical agent, in data and control communication with the monitoring system

where the analytical agent is operable to associate each metric in at least one set of metrics with at least one component and/or network device and obtain values for each such metric from the monitoring system, the analytical agent thereafter operable to determine whether each such metric is indicative of a problem with the data network and rank and correlate the indicative problems to determine whether a problem may be symptomatic of another problem based on an interconnection and/or interdependency between a physical machine and a virtual machine, between components or between components and network devices.

14. A computer readable medium having computer software recorded thereon such that, when the computer software is executed by a suitable processing system, the computer software is operable to:

associate each metric in a at least one set of metrics with at least one component and/or network device

obtain values for each such metric from a monitoring system;

determine whether each such metric is indicative of a problem within the data network; and

rank and correlate indicative problems to determine whether a problem may be symptomatic of another problem based on an interconnection and/or interdependency between a physical machine and a virtual machine, between components or between components and network devices.