METHODS AND DEVICES FOR DETERMINING QUALITY OF SERVICES OF STORAGE SYSTEMS

• US 20080244209A1
• Filed: 03/25/2008
• Published: 10/02/2008
• Est. Priority Date: 03/27/2007
• Status: Active Grant

• Current Assignee: Board of Regents of the University of Texas System (University of Texas System)
• Original Assignee: Board of Regents of the University of Texas System (University of Texas System)
• Inventors: Seelam, Seetharami R., Teller, Patricia J.
• Granted Patent: US 8,099,565 B2
• Time in Patent Office: Days
• US Class Current: 711/168
• CPC Class Codes: G06F 3/061 Improving I/O performance
• CPC Class Codes: G06F 3/0659 Command handling arrangements, e.g. command buffers, queues, command scheduling
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD
• CPC Class Codes: G06F 9/4881 Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues

Abstract

Methods and systems for allowing access to computer storage systems. Multiple requests from multiple applications can be received and processed efficiently to allow traffic from multiple customers to access the storage system concurrently.

Claims

1. A method for providing access to a storage system, the method comprising:

(a) receiving a plurality of requests from a plurality of applications for access to the storage system;

(b) assigning each application to one of a plurality of queues;

(c) calculating an initial quanta for each queue;

(d) selecting a first queue to access the storage system so that a request from the first queue can be serviced by the storage system;

(e) calculating a subsequent quanta for the first queue after the request from the first queue has been serviced by the storage system;

(f) determining if the subsequent quanta for the first queue is greater than or less than a predetermined value;

(g) selecting the first queue to access the storage system if the subsequent quanta for the first queue is greater than the predetermined value, so that a subsequent request from the first queue can be serviced by the storage system; and

(h) selecting a second queue to access the storage system if the subsequent quanta for the first queue is less than the predetermined value, so that a request from the second queue can be serviced by the storage system.

2. The method of claim 1 wherein steps (c) through (g) are repeated until the subsequent quanta for the first queue is less than the predetermined value.

3. The method of claim 2 further comprising:

(i) calculating a subsequent quanta for the second queue after the request from the second queue has been serviced by the storage system;

4. The method of claim 3 wherein the subsequent quanta for the first queue is less than the predetermined value by a specific amount in a round, and wherein the specific amount is subtracted from the subsequent quanta for the first queue in a subsequent round.

5. The method of claim 3 wherein the subsequent quanta for the second queue is less than the predetermined value by a specific amount in a round, and wherein the specific amount is subtracted from the subsequent quanta for the second queue in a subsequent round.

6. The method of claim 1 wherein the predetermined value is zero.

7. The method of claim 1 wherein each of the applications has a weight and the initial quanta calculated for each queue is based on the weight of the application.

8. The method of claim 1, where the plurality of requests for access to the storage system are received concurrently.

9. The method of claim 1, where allocating an available service time to each of the requests is controlled by a virtual input/output scheduler for coarse-grain allocation and a plurality of application-dependent schedulers for fine-grain allocation.

10. The method of claim 9, where the virtual input/output scheduler allocates service time to each of the requests in a round-robin fashion.

11. The method of claim 1, where a service time for the first queue to access the storage system does not affect the service time for the second queue to access the storage system.

12. A method comprising:

receiving a plurality of requests for access to the storage system;

receiving a plurality of priorities for each of the requesting applications;

calculating a quanta for each of a plurality of queues;

selecting a queue to access the storage system;

calculating the quanta left after the selected queue accesses the storage system;

selecting a next queue for access to storage; and

utilizing a virtual input/output scheduler and a plurality of application-dependent schedulers to allocate a first portion of the available service time to the first request and to allocate a second portion of the available service time to the second request.

13. The method of claim 12, where the first portion and the second portion are in proportion to an assigned weight for each of the queues.

14. The method of claim 12, where the first queue has a request to obtain a first amount of data, the second queue has a request to obtain a second amount of data, and each queue has its own application-specific scheduler.

15. The method of claim 12, where the virtual input/output scheduler allocates service to the first and second queues in a round-robin fashion.

16. The method of claim 12, where a service time calculation for the first request for access to the storage system does not affect a service time calculation for the second request for access to the storage system.

17. A method comprising:

receiving a plurality of requests from a plurality of applications for access to the storage system;

receiving a plurality of weights for each of the requesting applications;

assigning each application to one of a plurality of queues;

receiving from a first application a first request for storage access to a storage system;

receiving from a second application a second request for storage access to the storage system;

determining a bandwidth of the storage system;

determining a size for the first request;

determining a size for the second request;

determining a service time to satisfy the first request;

determining a service time to satisfy second request;

allocating a first portion of bandwidth to the first request; and

allocating a second portion of the bandwidth to the second request;

where the allocating of the first and second portions of bandwidth are based on the determined sizes and service times.

18. The method of claim 17, where the service time for the first queue for access to the storage system does not affect the service time for the second queue for access to the storage system.

19. The method of claim 17, where the allocated first portion and second portion of the service times are in proportion to the weights for the applications associated with the first request and the second request.

20. The method of claim 17, where the plurality of requests for access to the storage system are received concurrently and serviced concurrently.

21. The method of claim 17, where allocating the first portion and second portion of the service time is controlled by a virtual input/output scheduler for coarse-grain allocation and a plurality of application-dependent schedulers for fine-grain allocation.

22. The method of claim 17, where the virtual input/output scheduler allocates service time to the first and second requests in a round-robin fashion.