Processor provisioning by a middleware processing system

• US 9,298,508 B2
• Filed: 03/21/2014
• Issued: 03/29/2016
• Est. Priority Date: 12/21/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bird, Paul M., Kalmuk, David
• Primary Examiner(s): Truong, Lechi
• Application Number:
  US14/221,747
• Publication Number:
  US 20140201751A1
• Time in Patent Office: 739 Days
• Field of Search: 719/314
• US Class Current: 1/1
• CPC Class Codes: G06F 9/5016 the resource being the memory
• CPC Class Codes: G06F 9/5038 considering the execution order of a plurality of tasks, e.g. taking priority or time dependency constraints into consideration scheduling strategies G06F9/4881 and subgroups
• CPC Class Codes: G06F 9/5066 Algorithms for mapping a plurality of inter-dependent sub-tasks onto a plurality of physical CPUs mappping at compile time, see G06F8/451
• CPC Class Codes: G06F 9/546 Message passing systems or structures, e.g. queues

Abstract

A middleware processor provisioning process provisions a plurality of processors in a multi-processor environment. The processors themselves may be subdivided in to one or more partitions or processing instances for which a single processing queue is created and a single kernel thread is started. User processing requests are portioned and dispatched across the plurality of processing queues and are serviced by the corresponding kernel process, thereby efficiently using available processing resources while servicing the user processing requests in a desired manner.

Claims

1. A method implemented by a computer via a middleware software system for efficiently allocating workload for a plurality of user processing requests among a plurality of processors in a multiprocessor system comprising:

assessing a current processing workload for the plurality of user processing requests;

subdividing a processing capability of the multiprocessor system into a plurality of processing units based on the current processing workload;

starting one or more processes to service the user processing requests;

generating a processing queue for each of the processing units;

assigning a portion of each user processing request to one or more of the processing queues;

servicing the portion of each user processing request by the one or more processes;

allocating a variable predetermined time interval to each of the processing queues for which portions of the user processing requests assigned to a corresponding processing queue execute for a corresponding allocated time interval, wherein when a corresponding allocated time interval expires, a currently executing user processing request yields execution and signals the next in queue portion of the user processing requests to begin execution, and

when a currently executing user processing request finishes execution prior to the expiration of the corresponding allocated time interval, determining whether to yield execution of the corresponding allocated time interval or execute an additional portion of the user processing requests during the corresponding allocated time interval.

2. The method of claim 1, wherein said assigning is based on one or more of user processing request priority and user processing request service class.

3. The method of claim 2, further comprising collecting metrics associated with each of the user processing requests, wherein said processing request priority is based on the metrics;

and allocating a given portion of said processing capability of the multiprocessor system to a given user processing request service class when a user processing request service class is employed.

4. The method of claim 1, wherein a number of said plurality of processing units determines an associated concurrency level number for the multiprocessor system for which a corresponding number of processing queues are generated and said servicing includes concurrently servicing portions of the user processing requests at the front of each corresponding processing queue;

and

wherein when one of the portions of the user processing requests is executing any remaining portions of the user processing requests are blocked from execution and when said one of the portions of the user processing requests finishes executing or when a processing time allocated to a given processing queue associated with said one of the portions of the user processing requests expires a next scheduled portion of the user processing requests are unblocked from execution.

5. The method of claim 1, further comprising collecting metrics associated with each of the user processing requests and computing a moving average of the actual execution time associated with each variable predetermined time interval, and when the variable predetermined time interval drops below a threshold, reducing the sampling rate for collecting said metrics to an interval that is a multiple of a corresponding actual execution time.

6. The method of claim 1, further comprising determining when a portion of a user processing request is holding a resource needed by a higher priority user processing request;

and re-queuing the portion of the user processing request holding the resource to the front of one of the processing queues.

7. The method of claim 1, further comprising:

allocating a shared memory area to be shared by each of the processing queues;

dedicating a portion of the shared memory area to each of the processing queues; and

maintaining a counter by each of the processing queues in order for each of the processing queues to time execution of said user processing requests or portions thereof.