REDUCING READ LATENCY USING A POOL OF PROCESSING CORES

• US 20130339635A1
• Filed: 06/14/2012
• Published: 12/19/2013
• Est. Priority Date: 06/14/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: AMIT, Jonathan, LIDOR, Amir, MARENKOV, Sergey, RAIKHMAN, Rostislav
• Granted Patent: US 8,930,633 B2
• Time in Patent Office: Days
• US Class Current: 711/154
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3485 for I/O devices
• CPC Class Codes: G06F 2209/5011 Pool
• CPC Class Codes: G06F 2209/504 Resource capping
• CPC Class Codes: G06F 3/061 Improving I/O performance
• CPC Class Codes: G06F 3/0611 in relation to response time
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0631 by allocating resources to storage systems
• CPC Class Codes: G06F 3/0653 Monitoring storage devices or systems
• CPC Class Codes: G06F 3/0659 Command handling arrangements, e.g. command buffers, queues, command scheduling
• CPC Class Codes: G06F 3/0665 at area level, e.g. provisioning of virtual or logical volumes
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: G06F 3/0683 Plurality of storage devices
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

In a read processing storage system, using a pool of CPU cores, the CPU cores are assigned to process either write operations, read operations, and read and write operations, that are scheduled for processing. A maximum number of the CPU cores are set for processing only the read operations, thereby lowering a read latency. A minimal number of the CPU cores are allocated for processing the write operations, thereby increasing write latency. Upon reaching a throughput limit for the write operations that causes the minimal number of the plurality of CPU cores to reach a busy status, the minimal number of the plurality of CPU cores for processing the write operations is increased.

Claims

1. A method for reducing read latency in a read processing storage system using a pool of a plurality of CPU cores, comprising:

assigning to each one of the plurality of CPU cores for processing only one of write operations, read operations, and read and write operations that are scheduled for processing;

setting a maximum number of the plurality of CPU cores for processing only the read operations, thereby lowering the read latency; and

allocating a minimal number of the plurality of CPU cores for processing the write operations, thereby increasing a write latency, wherein upon reaching a throughput limit for the write operations that causes the minimal number of the plurality of CPU cores to reach a busy status, the minimal number of the plurality of CPU cores for processing the write operations is increased.

2. The method of claim 1, further including scheduling the write operations, the read operations, and the read and write operations, between the pool of the plurality of processing cores, according to a thread weight value (TWV) that is assigned to each one of the plurality of processing cores having a difference in processing power.

3. The method of claim 1, further including sustaining a bandwidth of the write operations by measuring core idle time and a throughput limit.

4. The method of claim 1, further including setting a time-out period for a predetermined threshold period for processing the write operations, the read operations, and the read and write operations.

5. The method of claim 1, further including assigning the read and write operations to each of the plurality of processing cores for processing if each of the plurality of CPU cores have attained the busy status.

6. The method of claim 1, wherein the busy status represents those of the plurality of CPU cores that have processed at least one of the write operations, the read operations, and the read and write operations during a predetermined threshold period.

7. The method of claim 1, further including assigning only the read operations to the plurality of CPU cores having an idle status, wherein the idle status represents those of the plurality of CPU cores that have yet to process any of the write operations, the read operations, and the read and write operations during a predetermined threshold period.

8. The method of claim 7, further including determining whether each one of the plurality of CPU cores had one of the idle status and the busy status during a last-predetermined threshold period.

9. The method of claim 1, further including determining the maximum number of the plurality of CPU cores by counting, from a firstmost positioned, one of the plurality of CPU cores, and sequentially counting to the right, until reaching a lastmost positioned one of the plurality of CPU cores, to include in the maximum number of the plurality of CPU cores to process the read operations.

10. The method of claim 9, further including determining the minimum number of the plurality of CPU cores by locating the lastmost positioned one of the plurality of CPU cores to process the write operations.

11. The method of claim 10, further including:

if the lastmost positioned one of the plurality of CPU cores is determined to be in the busy status, and those of the plurality of CPU cores to the left of the lastmost positioned one of the plurality of CPU cores are in the idle status;

permitting one of the plurality of CPU cores that is to the immediate left of the lastmost positioned one of the plurality of read to process the write operations.

12. The method of claim 11, further including:

if at least one or more of an nth number of lastmost positioned ones of the plurality of CPU cores are determined to be in the busy status, and all other of the plurality of CPU cores to the left of the at least one or more of the nth number of lastmost positioned ones of the plurality of CPU cores are in the idle status;

permitting an additional one of the at least one or more of the nth number of lastmost positioned ones of the plurality of CPU cores to process the write operations.

13. A system for reducing read latency in a read processing storage system using a pool of a plurality of CPU cores, comprising:

at least one scheduler scheduling write operations, read operations, and read and write operations between the pool of the plurality of processing cores,

at least one write cache, remotely connected to each of the plurality of CPU cores, assisting with the write operations;

at least one processor device, controlling the at least one scheduler and the at least one write cache, and operable in the read processing storage system, wherein the at least one processor device;

assigns to each one of the plurality of CPU cores for processing only one of write operations, read operations, and read and write operations that are scheduled for processing,

sets a maximum number of the plurality of CPU cores for processing only the read operations, thereby lowering the read latency, and

allocates a minimal number of the plurality of CPU cores for processing the write operations, thereby increasing a write latency, wherein upon reaching a throughput limit for the write operations that causes the minimal number of the plurality of CPU cores to reach a busy status, the minimal number of the plurality of CPU cores for processing the write operations is increased.

14. The system of claim 13, wherein the at least one processor device assigns a thread weight value (TWV) to each one of the plurality of processing cores having a difference in processing power, wherein the write operations, the read operations, and the read and write operations, between the pool of the plurality of processing cores, are proportionally scheduled by the scheduler according to the TWV.

15. The system of claim 13, wherein the at least one processor device sustains a bandwidth of the write operations by measuring core idle time and a throughput limit.

16. The system of claim 13, wherein the at least one processor device sets a time-out period for a predetermined threshold period for processing the write operations, the read operations, and the read and write operations.

17. The system of claim 13, wherein the at least one processor device assigns the read and write operations to each of the plurality of processing cores for processing if each of the plurality of CPU cores have attained the busy status.

18. The method of claim 13, wherein the busy status represents those of the plurality of CPU cores that have processed at least one of the write operations, the read operations, and the read and write operations during a predetermined threshold period.

19. The system of claim 13, wherein the at least one processor device assigns only the read operations to the plurality of CPU cores having an idle status, wherein the idle status represents those of the plurality of CPU cores that have yet to process any of the write operations, the read operations, and the read and write operations during a predetermined threshold period.

20. The system of claim 19, wherein the at least one processor device determines whether each one of the plurality of CPU cores had one of the idle status and the busy status during a last-predetermined threshold period.

21. The system of claim 13, wherein the at least one processor device determines the maximum number of the plurality of CPU cores by counting, from a firstmost positioned, one of the plurality of CPU cores, and sequentially counting to the right, until reaching a lastmost positioned one of the plurality of CPU cores, to include in the maximum number of the plurality of CPU cores to process the read operations.

22. The system of claim 21, wherein the at least one processor device determines the minimum number of the plurality of CPU cores by locating the lastmost positioned one of the plurality of CPU cores to process the write operations.

23. The system of claim 22, wherein the at least one processor device:

if the lastmost positioned one of the plurality of CPU cores is determined to be in the busy status, and those of the plurality of CPU cores to the left of the lastmost positioned one of the plurality of CPU cores are in the idle status;

permits one of the plurality of CPU cores that is to the immediate left of the lastmost positioned one of the plurality of read to process the write operations.

24. The system of claim 23, wherein the at least one processor device:

if at least one or more of an nth number of lastmost positioned ones of the plurality of CPU cores are determined to be in the busy status, and all other of the plurality of CPU cores to the left of the at least one or more of the nth number of lastmost positioned ones of the plurality of CPU cores are in the idle status;

permits an additional one of the at least one or more of the nth number of lastmost positioned ones of the plurality of CPU cores to process the write operations.

25. A computer program product for reducing read latency in a read processing storage system using a pool of a plurality of CPU cores, the computer program product comprising a computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:

a first executable portion that assigns to each one of the plurality of CPU cores for processing only one of write operations, read operations, and read and write operations that are scheduled for processing;

a second executable portion that sets a maximum number of the plurality of CPU cores for processing only the read operations, thereby lowering the read latency; and

a third executable portion that allocates a minimal number of the plurality of CPU cores for processing the write operations, thereby increasing a write latency, wherein upon reaching a throughput limit for the write operations that causes the minimal number of the plurality of CPU cores to reach a busy status, the minimal number of the plurality of CPU cores for processing the write operations is increased.

26. The computer program product of claim 25, further including a fourth executable portion that assigns a thread weight value (TWV) to each one of the plurality of processing cores having a difference in processing power, wherein the write operations, the read operations, and the read and write operations, between the pool of the plurality of processing cores, are proportionally scheduled according to the TWV.

27. The computer program product of claim 25, further including a fourth executable portion that sustains a bandwidth of the write operations by measuring core idle time and a throughput limit.

28. The computer program product of claim 25, further including a fourth executable portion that sets a time-out period for a predetermined threshold period for processing the write operations, the read operations, and the read and write operations.

29. The computer program product of claim 25, further including a fourth executable portion that assigns the read and write operations to each of the plurality of processing cores for processing if each of the plurality of CPU cores have attained the busy status.

30. The computer program product of claim 25, wherein the busy status represents those of the plurality of CPU cores that have processed at least one of the write operations, the read operations, and the read and write operations during a predetermined threshold period.

31. The computer program product of claim 25, further including a fourth executable portion that assigns only the read operations to the plurality of CPU cores having an idle status, wherein the idle status represents those of the plurality of CPU cores that have yet to process any of the write operations, the read operations, and the read and write operations during a predetermined threshold period.

32. The computer program product of claim 31, further including a fifth executable portion that determines whether each one of the plurality of CPU cores had one of the idle status and the busy status during a last-predetermined threshold period.

33. The computer program product of claim 25, further including a fourth executable portion that determines the maximum number of the plurality of CPU cores by counting, from a firstmost positioned, one of the plurality of CPU cores, and sequentially counting to the right, until reaching a lastmost positioned one of the plurality of CPU cores, to include in the maximum number of the plurality of CPU cores to process the read operations.

34. The computer program product of claim 33, further including a fifth executable portion that determines the minimum number of the plurality of CPU cores by locating the lastmost positioned one of the plurality of CPU cores to process the write operations.

35. The computer program product of claim 34, further including a sixth executable portion that:

if the lastmost positioned one of the plurality of CPU cores is determined to be in the busy status, and those of the plurality of CPU cores to the left of the lastmost positioned one of the plurality of CPU cores are in the idle status;

permits one of the plurality of CPU cores that is to the immediate left of the lastmost positioned one of the plurality of read to process the write operations.

36. The computer program product of claim 35, further including a seventh executable portion that:

if at least one or more of an nth number of lastmost positioned ones of the plurality of CPU cores are determined to be in the busy status, and all other of the plurality of CPU cores to the left of the at least one or more of the nth number of lastmost positioned ones of the plurality of CPU cores are in the idle status;

permits an additional one of the at least one or more of the nth number of lastmost positioned ones of the plurality of CPU cores to process the write operations.