Predictive bursty real-time traffic control for telecommunications switching systems

• US 6,252,950 B1
• Filed: 09/30/1998
• Issued: 06/26/2001
• Est. Priority Date: 09/30/1998
• Status: Expired due to Term

• Current Assignee: WSOU Investments, LLC (WSOU Holdings, LLC)
• Original Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Inventors: Ruiz, Jose Guadalupe, Duty, David Allen
• Primary Examiner(s): Kuntz, Curtis
• Assistant Examiner(s): TRAN, QUOC DUC
• Application Number:
  US09/163,614
• Time in Patent Office: 1,000 Days
• Field of Search: 379/111, 379/112, 379/113, 379/133, 379/268, 379/269, 379/137, 379/141, 379/288, 379/309
• US Class Current: 379/111
• CPC Class Codes: H04M 3/365 Load metering of control unit

Abstract

A traffic control system (and associated method) according to the present invention includes a feedback control mechanism in which the controlled variable is processor occupancy, the controlling variable is the amount of resources necessary to satisfy the operational requirements of the switching system, and the set point is the average expected load based on a Poisson arrival rate. Current processor occupancy is compared, on an instantaneous basis against the set point to determine if it exceeds the average expected load. If so, the controlling variable is adjusted to allocate additional resources to process the offered traffic load. The process continues until the traffic burst expires. Additional resources may be obtained by eliminating or deferring non-essential work, and by allocating additional call registers, effectively accepting work in larger increments or batches. By instantly and proactively responding to traffic bursts, this traffic control system aids the switching system in weathering brief periods of high traffic, without actually reaching an overload condition and invoking overload controls. As a result, switching system capacity is increased, and other performance characteristics are improved.

Claims

1. A traffic control for a telecommunications switching system, said system attained to receive offered calls and to accept some number of said offered calls for processing, said traffic control comprising:

means for measuring a performance value of said system indicative of the amount of system resources consumed in processing said accepted calls,

means for establishing a normal level of said performance value;

means for establishing a normal amount of said offered calls to be accepted for processing;

a plurality of call registers for storing information about said accepted calls, wherein each of said call registers is a data structure adapted to hold call-related information while processing a call;

each of said normal amount of accepted calls being respectively associated with at least one of said plurality of call registers during processing of said calls;

means for comparing said performance value to said normal level;

means responsive to said comparing means for causing said switching system to accept a number of calls greater than said normal amount when said performance value exceeds said normal level; and

means for allocating at least one additional call register for each of said number of accepted calls greater than said normal amount when said performance value exceeds said normal level.

2. The traffic control of claim 1 wherein:

said switching system is a stored-program control switching system comprising a processor, and some amount of time of said processor is allocated to processing calls.

3. The traffic control of claim 2, further comprising means responsive to said comparing means for causing said switching system to allocate additional processor time to call processing when said performance value exceeds said normal level.

4. The traffic control of claim 1 further comprising means responsive to said comparing means for deferring non-time-critical tasks when said performance value exceeds said normal level.

5. A traffic control for a telecommunications switching system, said system arranged to receive offered calls and during each of successive processing periods to accept some of said offered calls for processing, said traffic control comprising:

means for measuring a performance value of said system indicative of the amount of system resources consumed in processing said accepted calls during a previous processing period;

a plurality of call registers for storing information about said accepted calls, wherein each of said call registers is a data structure adapted to hold call-related information while processing a call;

said switching system respectively associating at least one of said call registers with each of said accepted calls;

means for establishing a normal level of said performance value;

means for establishing a normal amount of said call registers to be allocated for processing calls during a processing period;

means for comparing said performance value to said normal level; and

means responsive to said comparing means for causing said switching system to allocate an amount of said call registers greater than said normal amount when said performance value exceeds said normal level.

6. The traffic control of claim 5 wherein:

said switching system is a stored-program control switching system comprising a processor, and some amount of time of said processor is allocated to processing calls.

7. The traffic control of claim 6, further comprising means responsive to said comparing means for causing said switching system to allocate additional processor time to call processing when said performance value exceeds said normal level.

8. The traffic control of claim 5 further comprising means responsive to said comparing means for deferring non-time-critical tasks when said performance value exceeds said normal level.

9. The traffic control of claim 5 wherein switching system executes certain tasks during a base level cycle and said performance value is a measure of time required to complete said base level cycle.

10. A method for use with a telecommunications switching system, the switching system being arranged to receive offered calls and to accept some number of the offered calls for processing, the switching system having means for measuring a performance value indicative of the amount of system resources consumed in processing the accepted calls, comprising the steps of:

(a) establishing a normal level of said performance value;

(b) establishing a normal amount of said offered calls to be accepted for processing;

(c) accepting and processing a group of said offered calls;

(d) measuring said performance value resulting from said processing of said group of offered calls;

(e) comparing said performance value to said normal level;

(f) if said performance value does not exceed said normal level, then preparing to accept and process a next group of offered calls not larger than said established normal amount, including allocating a first number of call registers for use in processing said accepted calls, wherein each of said call registers is a data structure adapted to hold call-related information while processing a call;

(g) if said performance level does exceed said normal level, then preparing to accept and process a next group of offered calls larger than said established normal amount, including allocating a number of call registers larger than said first number for use in processing said accepted calls.

11. The method of claim 10 wherein said switching system is a stored-program switching system comprising a processor, and some amount of time of said processor is allocated to the function processing calls, and wherein step (g) thereof further comprises the step of:

(g1) allocating additional processor time to the function of processing calls.

12. The method of claim 10 wherein step (g) thereof further comprises the step of:

(g1) deferring non-time-critical tasks.

13. The method of claim 12 wherein step (g1) further comprises the steps of:

(g1a) determining an amount by which non-time-critical tasks were deferred during earlier processing of calls;

(g2a) if the amount by which non-time-critical tasks were deferred does not exceed a predetermined threshold, deferring non-time-critical tasks by a first amount during current or subsequent processing; and

(g3a) if the amount by which non-time-critical tasks were deferred exceeds a predetermined threshold, deferring non-time-critical tasks by a second amount smaller than said first amount during current or subsequent processing.

14. A method for use with a telecommunications switching system, the switching system being arranged to receive offered calls and during each of successive processing periods to accept some of the offered calls for processing, the switching system having means for measuring a performance value indicative of the amount of system resources consumed in processing the accepted calls, the switching system further having a plurality of information storage elements adapted to hold call-related information while processing a call, comprising the steps of:

(a) establishing a normal level of said performance value;

(b) establishing a normal amount of said offered calls to be accepted for processing;

(c) accepting and processing a group of said offered calls during a processing period;

(d) measuring said performance value resulting from said processing of said group of offered calls;

(e) comparing said performance value to said normal level;

(f) if said performance value does not exceed said normal level, then allocating for use in processing calls during a subsequent processing period a number of said information storage elements corresponding to that needed to process said normal amount of accepted calls, and accepting and processing a number of said offered calls no greater than said normal amount;

(g) if said performance level does exceed said normal level, then allocating for use in processing calls during a subsequent processing period a number of said information storage elements greater than that needed to process said normal amount of accepted calls, and accepting and processing a number of said offered calls greater than said normal amount.

15. The method of claim 14 wherein said switching system is a stored-program switching system comprising a processor, and some amount of time of said processor is allocated to the function processing calls, and wherein step (g) thereof further comprises the step of:

(g1) allocating additional processor time to the function of processing calls.

16. The method of claim 14 wherein step (g) thereof further comprises the step of:

(g1) deferring non-time-critical tasks.

17. The method of claim 16 wherein step (g) further comprises the steps of:

(g1a) determining an amount by which non-time-critical tasks were deferred during and earlier processing period;

(g2a) if the amount by which non-time-critical tasks were deferred does not exceed a predetermined threshold, deferring non-time-critical tasks by a first amount during a current or subsequent processing period; and

(g3a) if the amount by which non-time-critical tasks were deferred exceeds a predetermined threshold, deferring non-time-critical tasks by a second amount smaller than said first amount during a current or subsequent processing period.