Controlling allocation of system resources with an enhanced priority calculation

• US 6,584,488 B1
• Filed: 04/12/1999
• Issued: 06/24/2003
• Est. Priority Date: 04/12/1999
• Status: Expired due to Term

• Current Assignee: Trend Micro Inc.
• Original Assignee: International Business Machines Corporation
• Inventors: Brenner, Larry Bert, Browning, Luke Matthew
• Primary Examiner(s): Donaghue, Larry D
• Application Number:
  US09/289,885
• Time in Patent Office: 1,534 Days
• Field of Search: 709/103, 709/107
• US Class Current: 718/103
• CPC Class Codes: G06F 2209/5021 Priority
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals

Abstract

A method and system for controlling the allocation of a data processing system'"'"'s resources among two or more components competing for the resources. An internal system value is modified to yield a modified system value. A utilization history is created by mapping a component'"'"'s utilization of the system resources. Then, the priority of the component is calculated utilizing the utilization history and the modified system value within a priority calculation.

Claims

1. A method for controlling the allocation of a data processing system'"'"'s resources among two or more components competing for said resources, said method comprising the steps of:

adjusting a system value that is utilized as an offset for calculating a priority of said components when scheduling access to a specific system resource to yield a modified system value, wherein said modified system value accentuates the effect of said system value so that a resulting priority penalty is greater as the system value increases and said priority penalty is smaller as said system value decreases;

calculating the priority of said components utilizing said modified system value within a priority calculation, wherein said calculated priority is utilized to schedule access of said components to said specific system resource and is different from a priority calculated with said system value; and

scheduling said components according to said priority calculated utilizing said modified system value.

2. The method of claim 1, said adjusting step further including the step of selectively increasing said system value to obtain said modified system value, wherein when said component is desired to have a lower priority, said system value is dynamically increased.

3. The method of claim 1, wherein said component is a thread and said calculating step further comprises the steps of:

tracking a thread'"'"'s utilization of said resources to create a utilization history utilized to determine a correlated utilization value; and

calculating said priority of said thread utilizing said utilization value and said modified system value.

4. The method of claim 3, wherein said tracking step includes the step of monitoring a time period when said thread utilizes resources of a CPU of said data processing system.

5. The method of claim 3, wherein said data processing system contains a UNIX operating system and said system value is a NICE value, wherein further:

said tracking step includes the steps of monitoring said utilization value for a plurality of said threads; and

said calculating step includes the steps of;

assigning specific priorities to each of said plurality of said threads based on said utilization value and said modified system value; and

comparing said specific priorities of each of said plurality of threads to determine an effective priority of each of said threads.

6. The method of claim 5, wherein said calculating step yields a different priority based on said modified system value, wherein said calculating step further includes the step of changing a priority penalty for a unit of CPU time that a thread consumes, so that said penalty is greater as a NICE effect is increased and said penalty is smaller as said NICE effect is decreased, said priority penalty corresponding to said NICE value.

7. The method of claim 6, wherein said changing step includes the step of calculating said priority utilizing said modified system value within a priority equation, wherein said modified system value is dependent on a plurality of factors including:

level of priority desired for each of said plurality of threads;

length of time for sustaining a higher or lower priority once assigned;

utilization history of a particular thread; and

data processing system'"'"'s ability to efficiently handle multiple threads.

8. The method of claim 7, wherein when a thread is desired to be of low priority, said calculating step includes the step of increasing said NICE value to yield said modified system value.

9. The method of claim 8, wherein said increasing step includes the step of multiplying said NICE value by a factor greater than 1.

10. A system for controlling the allocation of a data processing system'"'"'s resources among two or more components competing for said resources, said system comprising:

means for adjusting a system value that is utilized as an offset for calculating a priority of said components when scheduling access to a specific system resource to yield a modified system value, wherein said modified system value accentuates the effect of said system value so that a resulting priority penalty is greater as the system value increases and said priority penalty is smaller as said system value decreases;

means for calculating the priority of said components utilizing said modified system value within a priority calculation, wherein said calculated priority is utilized to schedule access of said components to said specific system resource and is different from a priority calculated with said system value; and

means for scheduling said components according to said priority calculated utilizing said modified system value.

11. The system of claim 10, said adjusting means further including means for selectively increasing said system value to obtain said modified system values wherein when said component is desired to have a lower priority, said system value is dynamically increased.

12. The system of claim 10, wherein said component is a thread and said calculating means further comprises:

means for tracking a thread'"'"'s utilization of said resources to create a utilization history utilized to determine a correlated utilization value; and

means for calculating a priority of said thread utilizing said utilization value and said modified system value.

13. The system of claim 12, wherein said data processing system contains a UNIX operating system and said system value is a NICE value wherein further:

said tracking means includes means for monitoring said utilization value for a plurality of said threads; and

said calculating means includes means for;

assigning specific priorities to each of said plurality of said threads based on said utilization value and said modified NICE value; and

comparing said specific priorities of each of said plurality of threads to determine an effective priority of each of said threads.

14. The system of claim 13, wherein said calculating means yields a different priority based on said modified system value, wherein said calculating step further includes means for changing a priority penalty for a unit of CPU time that a thread consumes, so that said penalty is greater as a NICE effect is increased and said penalty is smaller as said NICE effect is decreased.

15. The system of claim 14, wherein said changing means includes means for calculating said priority utilizing said modified system value within a priority equation, wherein said modified system value is dependent on a plurality of factors including:

level of priority desired for each of said plurality of threads;

utilization history of a particular thread; and

the data processing system'"'"'s ability to efficiently handle multiple threads.

16. The system of claim 15, wherein when a thread is desired to be of low priority, said calculating means further includes means for multiplying said NICE value by a factor greater than 1 to yield said modified system value.

17. A computer program product for controlling the allocation of a data processing system'"'"'s resources among two or more components competing for said resources, said program product comprising:

a computer readable medium; and

program instructions on said computer readable medium for;

adjusting a system value that is utilized as an offset for calculating a priority of said components when scheduling access to a specific system resource to yield a modified system value, wherein said modified system value accentuates the effect of said system value so that a resulting priority penalty is greater as the system value increases and said priority penalty is smaller as said system value decreases;

calculating the priority of said components utilizing said modified system value within a priority calculation, wherein said calculated priority is utilized to schedule access of said components to said specific system resource and is different from a priority calculated with said system value; and

scheduling said components according to said priority calculated utilizing said modified system value.

18. The computer program product of claim 17, said program instructions for said adjusting step further including instructions for selectively increasing said system value to obtain said modified system value wherein when said component is desired to have a lower priority, said system value is dynamically increased.

19. The computer program product of claim 17, wherein said component is a thread, and said program instructions for said calculating step further includes instructions for:

tracking a thread'"'"'s utilization of said resources to create a utilization history utilized to determine a correlated utilization value; and

calculating a priority of said thread utilizing said utilization value and a modified system value.

20. The computer program product of claim 19, wherein said program instructions for said creating step includes program instructions for monitoring a time period when said thread utilizes resources of a CPU of said data processing system.

21. The computer program product of claim 19, wherein said program instructions are run on a data processing system containing a UNIX operating system and said system value is a NICE value, wherein further:

said program instructions for said tracking step includes program instructions for monitoring said utilization value for a plurality of said threads; and

said program instructions for said calculating step includes program instructions for;

assigning specific priorities to each of said plurality of said threads based on said utilization value and said modified NICE value; and

comparing said specific priorities of each of said plurality of threads to determine an effective priority of each of said threads.

22. The computer program product of claim 21, wherein said program instructions for said calculating step further includes program instructions for changing a priority penalty for a unit of CPU time that a thread consumes, wherein said penalty is greater as a NICE effect is increased and said penalty is smaller as said NICE effect is decreased.

23. The computer program product of claim 22, wherein said program instructions for said changing step includes program instructions for calculating said priority utilizing said NICE value within a priority equation, wherein said modified system value is dependent on a plurality of factors including:

level of priority desired for each of said plurality of threads;

utilization history of a particular thread; and

the data processing system'"'"'s ability to efficiently handle multiple threads.

24. The computer program product of claim 23, wherein when a thread is desired to be of low priority, said program instructions for said calculating step further includes program instructions for multiplying said NICE value by a factor greater than 1 to yield said modified system value.