Comprehensive bottleneck detection in a multi-tier enterprise storage system

• US 9,866,481 B2
• Filed: 05/07/2014
• Issued: 01/09/2018
• Est. Priority Date: 03/09/2011
• Status: Active Grant

• Current Assignee: Kyndryl Incorporated
• Original Assignee: International Business Machines Corporation
• Inventors: Agarwala, Sandip, Goodman, David P., Gopisetty, Sandeep, Weck, David D.
• Primary Examiner(s): Hussain, Imad
• Application Number:
  US14/271,572
• Publication Number:
  US 20140330961A1
• Time in Patent Office: 1,343 Days
• Field of Search: 709224
• CPC Class Codes: G06F 11/3442 for planning or managing the needed capacity
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 9/5072 Grid computing
• CPC Class Codes: H04L 41/147 for predicting network behaviour
• CPC Class Codes: H04L 43/0882 Utilisation of link capacity
• CPC Class Codes: H04L 47/11 Identifying congestion

Abstract

Embodiments of the present invention provide approaches (e.g., online methods) to analyze end-to-end performance issues in a multi-tier enterprise storage system (ESS), such as a storage cloud, where data may be distributed across multiple storage components. Specifically, performance and configuration data from different storage components (e.g., nodes) is collected and analyzed to identify nodes that are becoming (or may become) performance bottlenecks. In a typical embodiment, a set of components distributed among a set of tiers of an ESS is identified. For each component, a total capacity and a current load are determined. Based on these values, a utilization of each component is determined. Comparison of the utilization with a predetermined threshold and/or analysis of historical data allows one or more components causing a bottleneck to be identified.

Claims

1. A method for detecting a bottleneck in a multi-tier enterprise storage system (ESS), comprising:

identifying a set of components distributed among a set of tiers in the multi-tier ESS;

determining a total capacity of each of the set of components based on a corresponding set of component models;

determining a current load on each of the set of components;

determining, based on the total capacity and the current load, a utilization measurement for each of the set of components, the determining comprising an analysis of a utilization at a physical disk level, and an analysis of a utilization at at least one of a network shared disk level and a logical unit number level;

identifying at least one component, from the set of components, causing the bottleneck based on the utilization measurement;

determining whether the bottleneck is temporary by comparing a predicted load of the at least one component to the total capacity of the at least one component; and

providing remediation for the identified at least one component if the bottleneck is determined not to be temporary.

2. The method of claim 1, the ESS comprising a general parallel file system (GPFS).

3. The method of claim 1, the set or tiers comprising a file system tier, a storage server tier, a network shared disk (NSD) tier, a network tier, a logical unit number (LUN) tier, and a storage subsystem tier.

4. The method of claim 1, the set of tiers further comprising a client tier.

5. The method of claim 1, further comprising predicting a future bottleneck based on historical utilization data.

6. The method of claim 1, the utilization comprising a ratio of the current load to the total capacity, and the at least one component being identified based on a comparison of the utilization to a predetermined threshold.

7. The method of claim 1, further comprising computing an available bandwidth for each of the set of components by determining a difference between the total capacity and the current load.

8. The method of claim 1, wherein a service solution provider provides a computer infrastructure that performs the method for one or more consumers.

9. A system for detecting a bottleneck in a multi-tier enterprise storage system (ESS), comprising:

a bus;

a processor coupled to the bus; and

a memory medium coupled to the bus, the memory medium comprising instructions to;

identify a set of components distributed among a set of tiers in the multi-tier ESS;

determine a total capacity of each of the set of components based on a corresponding set of component models;

determine a current load on each of the set of components;

determine, based on the total capacity and the current load, a utilization measurement for each of the set of components, the determining comprising an analysis of a utilization at a physical disk level, and an analysis of a utilization at at least one of a network shared disk level and a logical unit number level;

identify at least one component, from the set of components, causing the bottleneck based on the utilization measurement;

determine whether the bottleneck is temporary by comparing a predicted load of the at least one component to the total capacity of the at least one component; and

provide remediation for the identified at least one component if the bottleneck is determined not to be temporary.

10. The system of claim 9, the ESS comprising a general parallel file system (GPFS).

11. The system of claim 9, the set or tiers comprising a file system tier, a storage server tier, a network shared disk (NSD) tier, a network tier, a logical unit number (LUN) tier, and a storage subsystem tier.

12. The system of claim 9, the set of tiers further comprising a client tier.

13. The system of claim 9, further comprising predicting a future bottleneck based on historical utilization data.

14. The system of claim 9, the utilization comprising a ratio of the current load to the total capacity, and the at least one component being identified based on a comparison of the utilization to a predetermined threshold.

15. A computer program product for detecting a bottleneck in a multi-tier enterprise storage system (ESS), the computer program product comprising a computer readable storage media, and program instructions stored on the computer readable storage media, to:

identify a set of components distributed among a set of tiers in the multi-tier ESS;

determine a total capacity of each of the set of components based on a corresponding set of component models;

determine a current load on each of the set of components;

determine, based on the total capacity and the current load, a utilization measurement for each of the set of components, the determining comprising an analysis of a utilization at a physical disk level, and an analysis of a utilization at at least one of a network shared disk level and a logical unit number level;

identify at least one component, from the set of components, causing the bottleneck based on the utilization measurement;

determine whether the bottleneck is temporary by comparing a predicted load of the at least one component to the total capacity of the at least one component; and

provide remediation for the identified at least one component if the bottleneck is determined not to be temporary.

16. The computer program product of claim 15, the ESS comprising a general parallel file system (GPFS).

17. The computer program product of claim 15, the set or tiers comprising a file system tier, a storage server tier, a network shared disk (NSD) tier, a network tier, a logical unit number (LUN) tier, and a storage subsystem tier.

18. The computer program product of claim 15, the set of tiers further comprising a client tier.

19. The computer program product of claim 15, further comprising predicting a future bottleneck based on historical utilization data.

20. The computer program product of claim 15, the utilization comprising a ratio of the current load to the total capacity, and the at least one component being identified based on a comparison of the utilization to a predetermined threshold.