REAL-TIME PERFORMANCE MODELING OF SOFTWARE SYSTEMS WITH MULTI-CLASS WORKLOAD

• US 20110087469A1
• Filed: 10/14/2009
• Published: 04/14/2011
• Est. Priority Date: 10/14/2009
• Status: Active Grant

• Current Assignee: Globalfoundries US Incorporated (GlobalFoundries, Inc.)
• Original Assignee: International Business Machines Corporation
• Inventors: Zhang, Li, Kumar, Dinesh, Tantawi, Asser N.
• Granted Patent: US 8,538,740 B2
• Time in Patent Office: Days
• US Class Current: 703/2
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis
• CPC Class Codes: G06F 2201/865 Monitoring of software

Abstract

Methods and systems for determining a real-time performance model of systems that process multi-class workloads. The methods can include collecting measurement data, selecting a series of prior time windows, processing the measurement data to compute a set of constraints based on the measurement data from the series of prior time windows, augmenting the set of constraints to a measurement model to obtain an augmented measurement model and running a modified extended Kalman filter with the augmented measurement model to obtain a state estimate.

Claims

1. In a performance modeling computer having a processor, a method for determining a real-time performance model of a system that processes multi-class workloads, the method comprising:

for a set of measurement data received by the processor;

collecting measurement datum at a time step k;

selecting N prior time windows;

processing the measurement datum to compute a set of constraints based on the measurement data from the N prior time windows;

augmenting the set of constraints to a measurement model to obtain an augmented measurement model;

running a modified extended Kalman filter with the augmented measurement model to obtain a state estimate; and

incrementing the time step k.

2. The method as claimed in claim 1 wherein the measurement model is given by:

z_k=H_kx_k+v_k

where z_k is a measurement vector for the measurement data, H_k is a measurement model matrix, x_k is a state vector and v_k is an observation noise vector.

3. The method as claimed in claim 2 wherein the set of constraints are augmented to the measurement model as measurements with zero noise covariance.

4. The method as claimed in claim 3 wherein the augmented measurement model is given by

5. The method as claimed in claim 2 wherein the set of constraints are augmented to the measurement model as measurements with non-zero noise covariance.

6. The method as claimed in claim 5 wherein the augmented measurement model is given by

7. The method as claimed in claim 2 wherein the set of constraints are given by

Dx_k=d

where D is a matrix and d is a vector.

8. The method as claimed in claim 7 where

d=[d₁d₂ . . . d_N]^T and D=[D₁D₂ . . . D_N]^T

9. The method as claimed in claim 8 further comprising processing the measurement vector over N time windows each of which is of size l_i sampling intervals.

10. The method as claimed in claim 9 wherein d_i is given by

11. The method as claimed in claim 10 wherein

12. The method as claimed in claim 9 wherein D_i is given by

13. The method as claimed in claim 12 wherein

14. In a performance modeling computer having a processor and a memory communicatively coupled to the processor, a method for determining a real-time performance model of a system that processes multi-class workloads, the method comprising:

collecting measurement data in the memory;

selecting a series of prior time windows;

processing the measurement data to compute a set of constraints based on the measurement data from the series of prior time windows;

augmenting the set of constraints to a measurement model to obtain an augmented measurement model; and

running a modified extended Kalman filter with the augmented measurement model to obtain a state estimate.

15. The method as claimed in claim 14 wherein the set of constraints is represented by constraint vectors and constraint matrices.

16. The method as claimed in claim 15 further comprising processing measurement vector data from the series of prior time windows to compute the constraint vectors.

17. The method as claimed in claim 14 further comprising processing measurement model matrices from the series of prior time windows to compute the constraint matrices.

18. A computer program product for determining a real-time performance model of a system that processes multi-class workloads, the computer program product including instructions for causing a performance modeling computer, having a processor and a memory communicatively coupled to the processor, to implement a method, the method comprising:

collecting measurement data in the memory;

selecting a series of prior time windows;

processing the measurement data to compute a set of constraints based on the measurement data from the series of prior time windows;

augmenting the set of constraints to a measurement model to obtain an augmented measurement model; and

running a modified extended Kalman filter with the augmented measurement model to obtain a state estimate.

19. The computer program product as claimed in claim 18 wherein the set of constraints is represented by constraint vectors and constraint matrices.

20. The computer program product as claimed in claim 19 further comprising:

processing measurement vector data from the series of prior time windows to compute the constraint vectors; and

processing measurement model matrices from the series of prior time windows to compute the constraint matrices.