Mitigating quantization effects in computer telemetry signals

1. A method for mitigating quantization effects in quantized telemetry signals, comprising:

monitoring a set of quantized telemetry signals;

for a given quantized telemetry signal in the set of quantized telemetry signals,

using a set of models to generate a set of estimates for the given quantized telemetry signal from the other monitored quantized telemetry signals, wherein each model in the set of models was initialized using a different randomly-selected subset of a training dataset, wherein the training dataset includes monitored telemetry signals from a computer system that is not degraded; and

averaging the set of estimates to produce an estimated signal for the given quantized telemetry signal.

2. The method of claim 1 , wherein prior to monitoring the set of quantized telemetry signals, the method further comprises:

receiving the training dataset for the set of quantized telemetry signals; and

for each quantized telemetry signal to be estimated, initializing the set of models for the quantized telemetry signal to be estimated based on a randomly-selected subset of the training dataset.

3. The method of claim 2 , wherein initializing a given model involves applying a non-linear, non-parametric regression technique to a randomly-selected subset of the training dataset.

4. The method of claim 3 , wherein after initializing the given model, the method further comprises returning the randomly-selected subset of the training dataset to the training dataset so that the randomly-selected subset of the training dataset can be reused when generating another model from the quantized telemetry signal.

5. The method of claim 3 , wherein the non-linear, non-parametric regression technique is a multivariate state estimation technique (MSET).

6. The method of claim 1 , wherein generating an estimate for the given quantized telemetry signal involves using a non-linear, non-parametric regression technique, which involves applying an associated model in the set of models to the other monitored quantized telemetry signals.

7. The method of claim 1 , wherein the method further comprises:

comparing one or more estimated signals with corresponding monitored signals to determine whether a second computer system is at the onset of degradation; and

if so, performing a remedial action.

8. A computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for mitigating quantization effects in quantized telemetry signals, wherein the method comprises:

monitoring a set of quantized telemetry signals;

for a given quantized telemetry signal in the set of quantized telemetry signals,

using a set of models to generate a set of estimates for the given quantized telemetry signal from the other monitored quantized telemetry signals, wherein each model in the set of models was initialized using a different randomly-selected subset of a training dataset, wherein the training dataset includes monitored telemetry signals from a computer system that is not degraded; and

averaging the set of estimates to produce an estimated signal for the given quantized telemetry signal.

9. The computer-readable storage medium of claim 8 , wherein prior to monitoring the set of quantized telemetry signals, the method further comprises:

receiving the training dataset for the set of quantized telemetry signals; and

for each quantized telemetry signal to be estimated, initializing the set of models for the quantized telemetry signal to be estimated based on a randomly-selected subset of the training dataset.

10. The computer-readable storage medium of claim 9 , wherein initializing a given model involves applying a non-linear, non-parametric regression technique to a randomly-selected subset of the training dataset.

11. The computer-readable storage medium of claim 10 , wherein after initializing the given model, the method further comprises returning the randomly-selected subset of the training dataset to the training dataset so that the randomly-selected subset of the training dataset can be reused when generating another model from the quantized telemetry signal.

12. The computer-readable storage medium of claim 10 , wherein the non-linear, non-parametric regression technique is a multivariate state estimation technique (MSET).

13. The computer-readable storage medium of claim 8 , wherein generating an estimate for the given quantized telemetry signal involves using a non-linear, non-parametric regression technique, which involves applying an associated model in the set of models to the other monitored quantized telemetry signals.

14. The computer-readable storage medium of claim 8 , wherein the method further comprises:

comparing one or more estimated signals with corresponding monitored signals to determine whether a second computer system is at the onset of degradation; and

if so, performing a remedial action.

15. An apparatus that mitigates quantization effects in quantized telemetry signals, comprising:

a processor;

memory coupled to the processor;

a monitoring mechanism configured to monitor a set of quantized telemetry signals;

an estimate-generation mechanism, wherein for a given quantized telemetry signal in the set of quantized telemetry signals, the estimate-generation mechanism is configured to use a set of models to generate a set of estimates for the given quantized telemetry signal from the other monitored quantized telemetry signals, wherein each model in the set of models was initialized using a different randomly selected subset of a training dataset, wherein the training dataset includes monitored telemetry signals from a computer system that is not degraded; and

an averaging mechanism configured to average the set of estimates to produce an estimated signal for the given quantized telemetry signal.

16. The apparatus of claim 15 , further comprising a model-generation mechanism, wherein prior to monitoring the set of quantized telemetry signals, the model-generation mechanism is configured to:

receive the training dataset for the set of quantized telemetry signals; and

for each quantized telemetry signal to be estimated, to initialize the set of models for the quantized telemetry signal to be estimated based on a randomly-selected subset of the training dataset.

17. The apparatus of claim 16 , wherein while initializing a given model, the model-generation mechanism is configured to apply a non-linear, non-parametric regression technique to a randomly-selected subset of the training dataset.

18. The apparatus of claim 17 , wherein after initializing the given model, the model-generation mechanism is configured to return the randomly-selected subset of the training dataset to the training dataset so that the randomly-selected subset of the training dataset can be reused when generating another model from the quantized telemetry signal.

19. The apparatus of claim 17 , wherein the non-linear, non-parametric regression technique is a multivariate state estimation technique (MSET).

20. The apparatus of claim 15 , wherein while generating an estimate for the given quantized telemetry signal, the estimate-generation mechanism is configured to use a non-linear, non-parametric regression technique, which involves applying an associated model in the set of models to the other monitored quantized telemetry signals.

21. The apparatus of claim 15 , further comprising a degradation-detection mechanism, wherein the degradation-detection mechanism is configured to:

compare one or more estimated signals with corresponding monitored signals to determine whether a second computer system is at the onset of degradation; and

if so, to perform a remedial action.