Workstation controller performance monitor

• US 4,905,171 A
• Filed: 11/09/1987
• Issued: 02/27/1990
• Est. Priority Date: 11/09/1987
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kiel, Harvey G., Van Heuklon, Jeffery J., Remfert, Jeffrey E.
• Primary Examiner(s): Lall, Parshotam S.
• Assistant Examiner(s): Cosimano, Edward R.
• Application Number:
  US07/117,864
• Time in Patent Office: 841 Days
• Field of Search: 364/200 MS File, 364/550, 364/551.01
• US Class Current: 702/186
• CPC Class Codes: G06F 11/3423 where the assessed time is active or idle time
• 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 11/3495 for systems
• CPC Class Codes: G06F 2201/81 Threshold
• CPC Class Codes: G06F 2201/88 Monitoring involving counting

Abstract

A workstation controller (WSC) performance monitor for locating performance bottlenecks is disclosed. The user notifies the host that he wants to start performance measurements. The host notifies each of the workstation controllers to begin collecting data that can be used to determine processor performance and communications line performance for each workstation controller. This data is collected in counters that reside in the storage of each of the workstation controllers. In the preferred embodiment, this data is periodically sent to a data base contained in the storage of the host. When the user specified length of time to collect data has expired, the host informs each of the connected workstation controllers to stop taking measurements. The data contained in the counters is sent to the database in the host and the host calculates processor utilization and communications line utilization for each of the connected workstation controllers. If the processor utilization falls below a first threshold, the performance of that processor is acceptable. If the processor utilization is greater than a second threshold, the performance of that processor is not acceptable. If the processor utilization is in between the first and second thresholds, the processor performance is considered marginal and a more complicated calculation of workload evenness of the processor is performed. If the workload is considered to be uneven, the performance of that processor is considered to be unacceptable.

Claims

1. A method for determining the performance of a processor in a computer system, comprising the steps of:

starting a performance test of said processor;

continuously checking to see if there are any tasks for said processor to perform;

incrementing an idle loop count if there are no tasks for said processor to perform, responsive to said checking step;

stopping said performance test after an elapsed time; and

determining a processor utilization using said elapsed time, an idle loop time, and said idle loop count, said idle loop time being the time said processor takes to perform one said checking step and one said incrementing step.

2. The method according to claim 1, further comprising the steps of:

indicating acceptable processor performance if said processor utilization is below a first threshold; and

indicating unacceptable processor performance if said processor utilization is above a second threshold.

3. The method according to claim 2, further comprising the steps of:

after said incrementing an idle loop count step.

periodically sampling a processor queue having tasks waiting to be performed;

incrementing a sample count responsive to said sampling step;

determining the number of tasks in said processor queue waiting to be performed;

incrementing a processor queue count by said number of tasks; and

after said stopping step.

determining a processor workload evenness by using said processor queue count and said sample count.

4. The method according to claim 3, further comprising the step of:

indicating uneven processor workload if said processor workload evenness exceeds a third threshold.

5. The method according to claim 4, further comprising the steps of:

indicating unacceptable processor performance if said processor utilization is between said first threshold and said second threshold and said processor workload evenness exceeds said third threshold.

6. A method for determining the performance of a communications line in a computer system, comprising the steps of:

starting a performance test of said communications line;

periodically sampling a communications queue;

incrementing a sample count responsive to said sampling step;

incrementing a communications usage count if there are tasks contained in said communications queue waiting to be performed;

stopping said performance test after an elapsed time;

determining a communications line utilization using said communications usage count and said sample count;

indicating acceptable communications line performance if said communications line utilization is below a first threshold; and

indicating unacceptable communications line performance if said communications line utilization is above a second threshold.

7. The method according to claim 6, further comprising the steps of:

determining the number of tasks in said communications queue waiting to be performed, responsive to said sampling step;

incrementing a communications queue count by said number of tasks;

determining a communications line workload evenness using said communications queue count and said sample count.

8. The method according to claim 7, further comprising the step of:

indicating uneven communications line workload if said communications line workload evenness exceeds a third threshold.

9. The method according to claim 8, further comprising the steps of:

indicating unacceptable communications line performance if said communications line utilization is between said first threshold and said second threshold and said communications line workload evenness exceeds said third threshold.

10. A computer system capable of diagnosing performance bottlenecks, comprising:

a plurality of workstation controllers each having a processor, each of said workstation controllers further comprising;

means for continuously checking to see if there are any tasks for said processor to perform;

means for incrementing an idle loop count if there are no tasks for said processor to perform, responsive to said checking step; and

a host computer, connected to said plurality of workstation controllers, further comprising;

means for determining said processor utilization by using an elapsed time, an idle loop time, and said idle loop count, said idle loop time being the time said processor takes to perform one said checking step and one said incrementing step.

11. The computer system of claim 10, wherein each of said workstation controllers further comprises:

means for indicating acceptable processor performance if said processor utilization is below a first threshold; and

means for indicating unacceptable processor performance if said processor utilization is above a second threshold.

12. The computer system of claim 11, wherein each of said workstation controllers further comprises:

means for periodically sampling a processor queue having tasks waiting to be performed;

means for incrementing a sample count responsive to said sampling means;

means for determining the number of tasks in said processor queue waiting to be performed;

means for incrementing a processor queue count by said number of tasks;

and wherein said host computer further comprises;

means for determining a processor workload evenness by using said processor queue count and said sample count.

13. The computer system of claim 12, wherein said host computer further comprises:

means for indicating uneven processor workload if said processor workload evenness exceeds a third threshold; and

means for indicating unacceptable processor performance if said processor utilization is between said first threshold and said second threshold and said processor workload evenness exceeds said third threshold.

14. The computer system of claim 13, wherein each of said workstation controllers further comprises:

means for periodically sampling a communications queue;

means for incrementing a communications usage count if there are tasks contained in said communications queue waiting to be performed;

and wherein said host computer further comprises;

means for determining a communications line utilization by using said communications usage count and said sample count.

15. The computer system of claim 14, wherein each of said workstation controllers further comprises:

means for indicating acceptable communications line performance if said communications line utilization is below a fourth threshold; and

means for indicating unacceptable communications line performance if said communications line utilization is above a fifth threshold.

16. The computer system of claim 15, wherein each of said workstation controllers further comprises:

means for determining the number of tasks in said communications queue waiting to be performed, responsive to said sampling means;

means for incrementing a communications line queue count by said number of tasks;

and wherein said host computer further comprises;

means for determining communications line workload evenness by using said communications line queue count and said sample count.

17. The computer system of claim 16, wherein said host computer further comprises:

means for indicating uneven communications line workload if said communications line workload evenness exceeds a sixth threshold; and

means for indicating unacceptable communications line performance if said communications line utilization is between said fourth threshold and said fifth threshold and said communications line workload evenness exceeds said sixth threshold.