Compression of telemetry data

1 . A non-transitory machine-readable storage medium comprising instructions that upon execution cause a system to:

select a slice size and a polynomial fitting order from a plurality of candidate slice sizes and a plurality of candidate polynomial fitting orders for representing a series of telemetry data obtained by sensors in a computing environment, wherein the selecting is based on solving an optimization problem comprising variables representing the slice size, the polynomial fitting order, and a measure of fit;

divide the series of telemetry data into a plurality of slices having the selected slice size;

apply compression of telemetry data in a slice of the plurality of slices using polynomial fitting according to the selected polynomial fitting order, the applied compression producing compressed telemetry data; and

cause transmission, from the system over a network, the compressed telemetry data to a processing system for processing of the compressed telemetry data.

2 . The non-transitory machine-readable storage medium of claim 1 , wherein the polynomial fitting is applied to first telemetry data in a first slice of the plurality of slices to produce a first compressed telemetry data portion, and the polynomial fitting is applied to second telemetry data in a second slice of the plurality of slices to produce a second compressed telemetry data portion.

3 . The non-transitory machine-readable storage medium of claim 2 , wherein the first compressed telemetry data portion comprises a first collection of polynomial coefficients produced by applying the polynomial fitting to the first telemetry data, and the second compressed telemetry data portion comprises a second collection of polynomial coefficients produced by applying the polynomial fitting to the second telemetry data.

4 . The non-transitory machine-readable storage medium of claim 3 , wherein a first polynomial fitting order selected for the first slice of the plurality of slices is different from a second polynomial fitting order selected for the second slice, and wherein a quantity of polynomial coefficients in the first collection of polynomial coefficients is different from a quantity of polynomial coefficients in the second collection of polynomial coefficients.

5 . The non-transitory machine-readable storage medium of claim 4 , wherein the instructions upon execution cause the system to:

select the first polynomial fitting order for the first slice by solving the optimization problem for the first slice; and

select the second polynomial fitting order for the second slice by solving the optimization problem for the second slice independently from solving the optimization problem for the first slice.

6 . The non-transitory machine-readable storage medium of claim 3 , wherein the instructions upon execution cause the system to:

reconstruct the first telemetry data at the processing system using the first collection of polynomial coefficients to produce reconstructed first telemetry data, and reconstruct the second telemetry data at the processing system using the second collection of polynomial coefficients to produce reconstructed second telemetry data.

7 . The non-transitory machine-readable storage medium of claim 6 , wherein the instructions upon execution cause the system to:

perform troubleshooting or management of the computing environment using reconstructed telemetry data comprising the reconstructed first telemetry data and the reconstructed second telemetry data.

8 . The non-transitory machine-readable storage medium of claim 1 , wherein the compressed telemetry data comprises polynomial coefficients produced by applying the polynomial fitting to the series of telemetry data, and wherein the instructions upon execution cause the system to:

apply bit quantization on the polynomial coefficients to produce quantized polynomial coefficients, wherein the compressed telemetry data comprises the quantized polynomial coefficients.

9 . The non-transitory machine-readable storage medium of claim 1 , wherein the polynomial fitting applied to telemetry data in the slice comprises a Chebyshev polynomial fitting.

10 . The non-transitory machine-readable storage medium of claim 1 , wherein the variables of the optimization problem comprise a first variable representing the slice size and the polynomial fitting order, and a second variable representing the measure of fit.

11 . The non-transitory machine-readable storage medium of claim 10 , wherein the first variable represents an aggregate of the slice size and the polynomial fitting order.

12 . The non-transitory machine-readable storage medium of claim 10 , wherein the instructions upon execution cause the system to:

represent values of the first variable using a first polynomial; and

represent values of the second variable using a second polynomial.

13 . The non-transitory machine-readable storage medium of claim 12 , wherein the solving of the optimization problem comprises finding an intersection of the first polynomial and the second polynomial.

14 . A computing system comprising:

a hardware processor; and

a non-transitory storage medium comprising instructions executable on the hardware processor to:

receive a series of telemetry data of a computing environment;

select a slice size and a polynomial fitting order from a plurality of candidate slice sizes and a plurality of candidate polynomial fitting orders for representing the series of telemetry data, wherein the selecting is based on solving an optimization problem comprising variables representing the slice size, the polynomial fitting order, and a measure of fit;

divide the series of telemetry data into a plurality of slices having the selected slice size;

apply polynomial fitting to telemetry data portions in respective slices of the plurality of slices, to produce respective compressed telemetry data portions, wherein the applying of the polynomial fitting comprises applying a polynomial fitting to a slice of the plurality of slices using the selected polynomial fitting order; and

cause transmission, from the computing system over a network, the compressed telemetry data portions to a processing system for processing of the compressed telemetry data portions.

15 . The computing system of claim 14 , wherein a first compressed telemetry data portion of the compressed telemetry data portions comprises a first collection of polynomial coefficients, and a second compressed telemetry data portion of the compressed telemetry data portions comprises a second collection of polynomial coefficients.

16 . The computing system of claim 14 , wherein the variables of the optimization problem comprise a first variable representing the slice size and the polynomial fitting order, and a second variable representing the measure of fit.

17 . The computing system of claim 16 , wherein the instructions are executable on the hardware processor to:

represent values of the first variable using a first polynomial; and

represent values of the second variable using a second polynomial.

18 . The computing system of claim 17 , wherein the solving of the optimization problem comprises finding an intersection of the first polynomial and the second polynomial.

19 . A method comprising:

receiving, at a computing system, a series of telemetry data obtained by one or more sensors of a computing environment;

selecting, by the computing system, a slice size and a polynomial fitting order from a plurality of candidate slice sizes and a plurality of candidate polynomial fitting orders for representing the series of telemetry data, wherein the selecting is based on solving an optimization problem comprising variables representing the slice size, the polynomial fitting order, and a measure of fit;

dividing, by the computing system, the series of telemetry data into a plurality of slices having the selected slice size;

applying, by the computing system, polynomial fitting according to the selected polynomial fitting order to telemetry data portions in respective slices of the plurality of slices, to produce respective compressed telemetry data portions; and

transmitting, from the computing system over a network, the compressed telemetry data portions to a processing system for processing of the compressed telemetry data portions.

20 . The method of claim 19 , further comprising:

reconstructing, at the processing system, telemetry data from polynomial coefficients included in the compressed telemetry data portions.