Analytical bandwidth enhancement for monitoring telemetric signals

1. A computer-implemented method for analyzing telemetry data from a monitored system, comprising:

in one or more computers, performing operations for:

obtaining the telemetry data as a set of telemetric signals from the monitored system;

grouping the telemetry data into one or more clusters of correlated telemetric signals from the telemetric signals; and

increasing a bandwidth associated with monitoring the telemetric signals by:

omitting one or more of the correlated telemetric signals from each of the clusters during sampling of the telemetric signals; and

estimating the omitted correlated telemetric signals from the sampled telemetric signals.

2. The computer-implemented method of claim 1 , wherein grouping the telemetry data into one or more clusters of correlated telemetric signals involves:

identifying a cross-predictability measure associated with each telemetric signal from the telemetric signals using a cross-correlation analysis technique; and

creating the clusters based on the cross-predictability measure.

3. The computer-implemented method of claim 2 , wherein the cross-predictability measure corresponds to a prediction error associated with the telemetric signal.

4. The computer-implemented method of claim 3 , wherein the telemetric signal is continuously sampled at a high sampling rate if the prediction error is high.

5. The computer-implemented method of claim 1 , wherein the monitored system corresponds to a computer system.

6. The computer-implemented method of claim 5 , wherein the telemetric signals from the computer system comprise at least one of:

a load metric;

a CPU utilization;

an idle time;

a memory utilization;

a disk activity;

a transaction latency;

a temperature;

a voltage;

a fan speed; and

a current.

7. The computer-implemented method of claim 1 , wherein estimating the omitted correlated telemetric signals involves:

applying a nonlinear, nonparametric regression technique to the sampled telemetric signals.

8. The computer-implemented method of claim 1 , wherein the nonlinear, nonparametric regression technique corresponds to a multivariate state estimation technique (MSET).

9. A system for analyzing telemetry data from a monitored system, comprising:

a monitoring mechanism configured to obtain the telemetry data as a set of telemetric signals from the monitored system;

a cross-correlation apparatus configured to group the telemetry data into one or more clusters of correlated telemetric signals from the telemetric signals; and

an analysis apparatus configured to increase a bandwidth associated with monitoring the telemetric signals by:

omitting one or more of the correlated telemetric signals from each of the clusters during sampling of the telemetric signals by the monitoring mechanism; and

estimating the omitted correlated telemetric signals from the sampled telemetric signals.

10. The system of claim 9 , wherein grouping the telemetry data into one or more clusters of correlated telemetric signals involves:

identifying a cross-predictability measure associated with each telemetric signal from the telemetric signals using a cross-correlation analysis technique; and

creating the clusters based on the cross-predictability measure.

11. The system of claim 10 , wherein the cross-predictability measure corresponds to a prediction error associated with the telemetric signal.

12. The system of claim 11 , wherein the telemetric signal is continuously sampled at a high sampling rate if the prediction error is high.

13. The system of claim 9 , wherein estimating the omitted correlated telemetric signals involves:

applying a nonlinear, nonparametric regression technique to the sampled telemetric signals.

14. The system of claim 9 , wherein the nonlinear, nonparametric regression technique corresponds to a multivariate state estimation technique (MSET).

15. A non-transitory computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for analyzing telemetry data from a monitored system, the method comprising:

obtaining the telemetry data as a set of telemetric signals from the monitored system;

grouping the telemetry data into one or more clusters of correlated telemetric signals from the telemetric signals; and

increasing a bandwidth associated with monitoring the telemetric signals by:

omitting one or more of the correlated telemetric signals from each of the clusters during sampling of the telemetric signals; and

estimating the omitted correlated telemetric signals from the sampled telemetric signals.

16. The non-transitory computer-readable storage medium of claim 15 , wherein grouping the telemetry data into one or more clusters of correlated telemetric signals involves:

identifying a cross-predictability measure associated with each telemetric signal from the telemetric signals using a cross-correlation analysis technique; and

creating the clusters based on the cross-predictability measure.

17. The non-transitory computer-readable storage medium of claim 16 , wherein the cross-predictability measure corresponds to a prediction error associated with the telemetric signal.

18. The non-transitory computer-readable storage medium of claim 17 , wherein the telemetric signal is continuously sampled at a high sampling rate if the prediction error is high.

19. The non-transitory computer-readable storage medium of claim 15 , wherein estimating the omitted correlated telemetric signals involves:

applying a nonlinear, nonparametric regression technique to the sampled telemetric signals.

20. The non-transitory computer-readable storage medium of claim 15 , wherein the nonlinear, nonparametric regression technique corresponds to a multivariate state estimation technique (MSET).