Adaptive Scheduling Of Storage Operations Based On Utilization Of Multiple Client And Server Resources In A Distributed Network Storage System

• US 20100076805A1
• Filed: 09/24/2008
• Published: 03/25/2010
• Est. Priority Date: 09/24/2008
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Batsakis, Alexandros, Talpey, Thomas, Kanevsky, Arkady, Lentini, James
• Granted Patent: US 8,392,312 B2
• Time in Patent Office: Days
• US Class Current: 705/8
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]
• CPC Class Codes: G06Q 30/08 Auctions
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• 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]

Abstract

Scheduling operations such as asynchronous file system operations in a network storage system is accomplished by applying a bid-price online auction methodology, in which bid (willingness-to-pay) values and price (cost) values are dynamically set by storage clients and a storage server, respectively, based on utilization of computing resources. The system provides a framework for adaptively scheduling asynchronous file system operations, managing multiple key resources of the distributed file system, including network bandwidth, server I/O, server CPU, and client and server memory utilization. The system can accelerate, defer, or cancel asynchronous requests to improve application-perceived performance. Congestion pricing via online auctions can be employed to coordinate the use of system resources by clients, so clients can detect shortages and adapt their resource usage.

Claims

1. A method comprising:

operating a network storage system that includes a storage server and a plurality of storage clients; and

scheduling storage operations in the network storage system by applying a bid-price auction model, wherein bid values and price values are dynamically set by the storage clients and the storage server, respectively, based on utilization of computing resources of each of the storage clients and the storage server.

2. A method as recited in claim 1, wherein scheduling storage operations comprises:

scheduling execution of file system operations by applying the bid-price auction model.

3. A method as recited in claim 1, wherein scheduling storage operations comprises:

scheduling execution of asynchronous file system operations by applying the bid-price auction model.

4. A method as recited in claim 1, wherein each of the clients dynamically sets a bid value based on current utilization of computing resources of that storage client, and the storage server dynamically sets a price value based on current utilization of computing resources of the storage server.

5. A method as recited in claim 4, wherein scheduling storage operations comprises:

for each of a plurality of storage operation requests from one or more of the storage clients, destined for the storage server, determining a disposition of the request based on a relationship between a bid value and a price value.

6. A method of operating a storage client in a network storage system that includes a storage server and the storage client, the method comprising:

determining in the storage client a current utilization of computing resources of the storage client;

computing a bid value in the storage client based on the current utilization of the computing resource of the storage client;

receiving a price value from the storage server, the price value having been computed by the storage server based on a current utilization of computing resources of the storage server;

generating a request in the storage client for the storage server to perform a storage operation; and

determining a disposition of the request based on a relationship between the bid value and the price value.

7. A method as recited in claim 6, wherein the storage operation is an asynchronous write, and wherein determining a disposition of the request comprises:

deferring sending the request to the storage server when the bid value is less than the price value.

8. A method as recited in claim 6, wherein the storage operation is an asynchronous write, and wherein determining a disposition of the request comprises:

accelerating execution of the write when the bid value is greater than the price value.

9. A method as recited in claim 6, wherein the storage operation is an asynchronous read, and wherein determining a disposition of the request comprises:

canceling the request when the bid value is less than the price value.

10. A method as recited in claim 6, wherein computing the bid value comprises computing the bid value based on utilization of memory of the storage client;

wherein the price value has been computed based on utilization of CPU and memory of the storage server.

11. A method as recited in claim 10, wherein the price value further has been computed based on:

utilization of network bandwidth; and

utilization of nonvolatile storage of the storage server.

12. A method as recited in claim 6, further comprising:

requesting the price value from the storage server.

13. A method of operating a network storage system that includes a storage server and a plurality of storage clients, the method comprising:

in the storage server,

determining a current utilization of computing resources of the storage server,

computing a price value based on the current utilization of computing resources of the storage server, and

sending the price value to each of the storage clients; and

in each of the storage clients,

determining a current utilization of computing resources of the storage client,

computing a bid value for an asynchronous file system operation, based on the current utilization of the computing resource of the storage client,

generating a request for the storage server to perform the asynchronous filesystem operation, and

determining a disposition of the request based on a relationship between the bid value and the price value, wherein

when the asynchronous filesystem operation is an asynchronous write, the storage client defers sending the request to the storage server if the bid value is less than the price value and accelerates execution of the write if the bid value is greater than the price value, and

when the storage operation is an asynchronous read, the storage client cancels the request if the bid value is less than the price value.

14. A method as recited in claim 13, wherein:

the bid value is computed by each client based on utilization of memory of the storage client; and

the price value is computed based on utilization of CPU and memory of the storage server.

15. A method as recited in claim 14, wherein the price value further is computed based on:

utilization of network bandwidth; and

utilization of nonvolatile storage of the storage server.

16. A processing system configured to operate as a network endpoint in a network storage system, the processing system comprising:

a processor;

a network interface through which to communicate with another network endpoint via a network; and

a storage medium storing instructions which, when executed by the processor, cause the processing system to perform a process that includes;

determining a current utilization of a computing resource of the processing system;

computing a first value based on the current utilization of the computing resource of the processing system; and

determining a disposition of a request for a storage operation in the network storage system, based on a relationship between the first value and a second value, the second value having been computed based on a current utilization of resources of the network storage system that are outside said processing system.

17. A processing system as recited in claim 16, wherein the processing system is a storage client, and wherein the second value is based on current utilization of resources of a storage server in the network storage system.

18. A processing system as recited in claim 17, wherein said process further comprises:

requesting the second value from the storage server;

receiving the second value from the storage server; and

generating the request for the storage operation.

19. A processing system as recited in claim 16, wherein the storage operation is a file system operation.

20. A processing system as recited in claim 16, wherein the storage operation is an asynchronous file system operation.

21. A processing system as recited in claim 20, wherein the storage operation is an asynchronous write, and wherein determining a disposition of the request comprises:

deferring sending the request to the storage server when the bid value is less than the price value.

22. A processing system as recited in claim 20, wherein the storage operation is an asynchronous write, and wherein determining a disposition of the request comprises:

accelerating execution of the write when the bid value is greater than the price value.

23. A processing system as recited in claim 20, wherein the storage operation is an asynchronous read, and wherein determining a disposition of the request comprises:

canceling the request when the bid value is less than the price value.

24. A processing system as recited in claim 16, wherein:

the first value is based on utilization of memory of the processing system; and

the second value is based on utilization of CPU and memory of another processing system in the network storage system.

25. A processing system as recited in claim 24, wherein the second value further is based on:

utilization of network bandwidth; and

utilization of nonvolatile storage of said other processing system.

26. A system comprising:

a network endpoint in a network storage system that includes a storage server and a storage client, the network endpoint being either the storage server or the storage client; and

scheduling means for scheduling storage operations in the network storage system by applying a bid-price auction model, wherein bid values and price values are dynamically set by the storage client and the storage server, respectively, based on utilization of computing resources of the storage client and the storage server.

27. A system as recited in claim 26, wherein the scheduling means comprises:

means for scheduling execution of file system operations by applying the bid-price auction model.

28. A system as recited in claim 26, wherein the scheduling means comprises:

means for scheduling execution of asynchronous file system operations by applying the bid-price auction model.

29. A system as recited in claim 26, wherein the scheduling means comprises:

means for dynamically setting a bid value based on current utilization of computing resources of the storage client, and

means for dynamically setting a price value based on current utilization of computing resources of the storage server.

30. A system as recited in claim 29, wherein the scheduling means further comprises means for determining a disposition of a storage operation request based on a relationship between the bid value and the price value.