Proportional quality of service based on client usage and system metrics

• US 9,838,269 B2
• Filed: 04/04/2013
• Issued: 12/05/2017
• Est. Priority Date: 12/27/2011
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Wright, David D., Xu, Michael
• Primary Examiner(s): BILGRAMI, ASGHAR H
• Application Number:
  US13/856,958
• Publication Number:
  US 20130227111A1
• Time in Patent Office: 1,706 Days
• Field of Search: 703223, 703224
• CPC Class Codes: G06F 11/3433 for load management allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029
• CPC Class Codes: G06F 11/3485 for I/O devices
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system
• CPC Class Codes: H04L 41/50 Network service management, e.g. ensuring proper service fulfilment according to agreements
• CPC Class Codes: H04L 41/5025 by proactively reacting to service quality change, e.g. by reconfiguration after service quality degradation or upgrade
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34
• CPC Class Codes: H04L 67/61 taking into account QoS or priority requirements

Abstract

Disclosed are systems, computer-readable mediums, and methods for determining client metrics of a volume in a storage system for a first client of a plurality of clients. The storage system stores data from the plurality of clients. System metrics of a cluster in the storage system are determined based upon use of the storage system by the plurality of clients. A load value of the storage system is determined based upon the system metrics and the client metrics. The load value is determined to be above a predefined threshold. A target performance value is calculated based upon the load value, a minimum quality of service value, and a maximum quality of service value. Performance of the storage system is adjusted for the client based upon the target performance value and the determining the load value is above the predefined threshold.

Claims

1. A method comprising:

determining a system load value of a distributed storage system based upon a plurality of service load values calculated for services on nodes of the distributed storage system;

determining the system load value is above a predefined threshold;

calculating proportional contribution to the system load value for each of a plurality of clients based on the plurality of service load values calculated for the services;

identifying at least a first client of the plurality of clients for performance throttling based on the calculated proportional contributions;

calculating a target performance value for the first client, wherein calculation of the target performance value is based upon the calculated proportional contribution of the first client to the system load value and quality of service parameter values for the first client; and

throttling performance of the distributed storage system for the first client based upon the target performance value.

2. The method of claim 1, wherein the plurality of service load values comprises write latency load values of nodes of the distributed storage system, wherein the system load value is the maximum value of write latency load values associated with two different nodes, wherein the two different nodes are both associated with the first client.

3. The method of claim 1, wherein the plurality of service load values comprises at least one of a read latency or a write latency of one or mores services on the nodes.

4. The method of claim 1, wherein calculating the proportional contribution of the first client comprises calculating a ratio of those of the plurality of service load values associated with the first client and a total of the plurality of service load values.

5. The method of claim 1, further comprising determining a class of the first client from a plurality of classes and determining the quality of service parameters for the first client based upon the class of the first client.

9. The method of claim 5, wherein identifying the first client for performance throttling is also based on the class of the first client as compared to a class of a second client of the plurality of clients.

10. The method of claim 1, wherein the plurality of service load values are based on at least one of read load values, write load values, write buffer load values, and a combination of at least two of the different types of load values.

11. The method of claim 10 further comprising determining which type of load values represents a greatest load on the distributed storage system and calculating the plurality of service load values with the load values of the type that represents the greatest load.

12. The method of claim 11, wherein determining which type of load values represents a greatest load on the distributed storage system is for each of the plurality of clients.

13. The method of claim 1, wherein the quality of service parameter values comprise a minimum quality of service value, a maximum quality of service value, and a maximum burst value, wherein throttling the performance is also based on a difference between the minimum and maximum quality of service values of the first client.

6. A non-transitory computer-readable medium having instructions stored thereon, that when executed by a computing device cause the computing device to perform operations comprising:

determining a system load value of a distributed storage system based upon a plurality of service load values calculated for services on nodes of the distributed storage system;

determining the system load value is above a predefined threshold;

calculating a proportional contribution to the system load value for each of a plurality of clients based on the plurality of service load values calculated for the services;

identifying at least a first client of the plurality of clients for performance throttling based on the calculated proportional contributions;

calculating a target performance value for the first client, wherein calculation of the target performance value is based upon the calculated proportional contribution of the first client to the system load value and quality of service parameters of the first client; and

throttling performance of the distributed storage system for the first client based upon the target performance value.

19. The non-transitory computer-readable medium of claim 6, wherein the operations further comprise determining class of the first client, wherein identifying the first client for performance throttling comprises identifying the first client also based on the class of the first client as compared to a class of a second client of the plurality of clients.

20. The non-transitory computer-readable medium of claim 6, wherein the quality of service parameter values comprise a minimum quality of service value, a maximum quality of service value, and a maximum burst value, wherein throttling the performance comprises throttling also based on a difference between the minimum and maximum quality of service values of the first client.

7. A system comprising:

a processor;

a network interface;

a non-transitory computer-readable medium having instructions executable by the processor to cause the system to, determine a system load value of a distributed storage system based upon a plurality of service load values calculated for services on nodes of the distributed storage system;

determine the system load value is above a predefined threshold;

calculate proportional contribution to the system load value for each of a plurality of clients based on the plurality of service load values calculated for the services;

identify at least a first client of the plurality of clients for performance throttling based on the calculated proportional contributions;

calculate a target performance value based upon the system load value and quality of service values of the first client; and

throttle performance of the distributed storage system for the first client based upon the target performance value.

8. The system of claim 7, wherein the computer-readable medium further comprises instructions executable by the processor to cause the system to:

normalize one or more system metrics based upon a performance curve and an input/output data size of input/output operations associated with the storage system; and

normalize one or more client metrics based upon a performance curve and an input/output data size of input/output operations associated with the first client.

14. The system of claim 7, wherein the computer-readable medium further comprises instructions executable by the processor to cause the system to determine class of the first client, wherein the instructions to identify the first client for performance throttling are executable by the processor to cause the system to identify the first client also based on the class of the first client as compared to a class of a second client of the plurality of clients.

15. The system of claim 7, wherein the plurality of service load values are based on at least one of read load values, write load values, write buffer load values, and a combination of at least two of the different types of load values.

16. The system of claim 15, wherein the computer-readable medium further comprises instructions executable by the processor to cause the system to determine which type of load values represents a greatest load on the distributed storage system and to calculate the plurality of service load values with the load values of the type that represents the greatest load.

17. The system of claim 16, wherein the instructions to determine which type of load values represents a greatest load on the distributed storage system comprises the instructions executable by the processor to determine, for each of the plurality of clients, which type of load values associated with the client represents a greatest load on the distributed storage system.

18. The system of claim 7, wherein the quality of service parameter values comprise a minimum quality of service value, a maximum quality of service value, and a maximum burst value, wherein the instructions to cause the system to throttle the performance comprises the instructions executable by the processor to cause the system to throttle also based on a difference between the minimum and maximum quality of service values of the first client.