Frequency-domain analysis of data-center operational and performance metrics

Patent metadata

Field	Value
Publication number	US-11245593-B2
Application number	US-201615138102-A
Country	US
Kind code	B2
Filing date	Apr 25, 2016
Priority date	Apr 25, 2016
Publication date	Feb 8, 2022
Grant date	Feb 8, 2022

Abstract

The current document is directed to methods and systems for frequency-domain analysis of operational and performance metric values and other data generated and collected within computer systems, including large distributed computer systems and virtualized data centers. In one implementation, each set of time-ordered values for each metric in a set of metrics is partitioned into time intervals, transformed from the time domain to the frequency domain, and aligned to generate a metric surface in a frequency-time-amplitude space. The metric surfaces are then pairwise compared to identify related metrics. Transfer functions are generated for transforming metric surfaces into one another. The comparison values and transfer functions are used to produce graphs that encapsulate discovered relationships between metrics.

First claim

The invention claimed is:

1. A method, carried out in a computer system having one or more processors, one or more memories, and one or more mass-storage devices, that compares a first sequence of time-ordered metric values for a particular metric generated by a first component of a computer system to a second sequence of time-ordered metric values for the particular metric generated by a second component of the computer system to produce a comparison metric that reflects the similarity of the operational behaviors of the different components of the computer system, the method comprising:
selecting a total time span that includes a sequence of pairs of time-aligned metric values, each pair included a first metric value selected from the first sequence of metric values and a second metric value selected from the second sequence of metric values;
dividing the total time span into multiple time intervals;
for each time interval of the multiple time intervals,
generating, from a portion the first sequence of time-ordered metric values within the time interval, representing time-domain data, a set of frequency-domain data points for the time interval, and
generating, from a portion the second sequence of time-ordered metric values within the time interval, representing two-dimensional time-domain data with time and metric-value axes, a set of frequency-domain data points for the time interval, representing two-dimensional frequency-domain data with frequency and amplitude axes;

stacking the sets of frequency-domain data points generated from portions of the first sequence of time-ordered metric values along a time axis to create a first three-dimensional volume with frequency, time, and amplitude axes;
stacking the sets of frequency-domain data points generated from portions of the second sequence of time-ordered metric values along a time axis to create a second three-dimensional volume with frequency, time, and amplitude axes;
interpolating a first three-dimensional metric surface from the first three-dimensional volume which represents mixed time-domain and frequency-domain data;
interpolating a second three-dimensional metric surface from the first three-dimensional volume which represents mixed time-domain and frequency-domain data;
aligning the first and second metric surfaces in three dimensions using a set of one or more surface-modifying operations:
generating the comparison metric from the aligned first and second metric surfaces; and
using the comparison metric to diagnose and ameliorate problematic operational states of the computer system and/or to tuning the computer system to increase performance of the computer system.

2. The method of claim 1 wherein generating the comparison metric from the aligned first and second metric surfaces further comprises:
generating an initial comparison metric value in; and
generating the comparison metric as a function of m and weighted sums of discount values associated with each of the surface-modifying operations used to align the first and second metric surfaces.

3. The method of claim 2 wherein the initial comparison metric value m is generated by one of:
summing the volumes of volume elements between the aligned first and second metric surfaces; and
comparing metrics associated with common features of the first and second metric surfaces.

4. The method of claim 1
wherein the first and second sequences of time-ordered metric values each comprises time/metric-value data points that represent sampled values of the particular metric; and
wherein the particular metric is selected from among a performance metric, an operational-characteristic metric, and an operational-behavior metric for the computer system.

5. The method of claim 1 wherein the particular metric is related to the performance, operational characteristics, or operational behavior of one of:
the computer system;
a single component or subcomponent of the computer system; and
an aggregation of multiple components and/or subcomponents of the computer system.

6. The method of claim 1 wherein the particular metric is related to values computed from one or more sets of time-ordered metric values related to the performance, operational characteristics, or operational behavior of one of:
the computer system;
a single component or subcomponent of the computer system; and
an aggregation of multiple components and/or subcomponents of the computer system.

7. The method of claim 1 wherein the computer system is one of:
a personal computer;
a server;
a workstation;
a multi-computer-system data center;
a distributed computer system;
a virtual data center;
a geographically distributed computer system; and
and aggregation of geographically distributed data centers.

8. The method of claim 1 wherein a set of frequency-domain data points for the time interval are generated from a portion of a sequence of time-ordered metric values within the time interval by application of a discrete Fourier transform.

9. The method of claim 1 wherein the surface-modifying operations used to align the first and second metric surfaces in three dimensions are selected from among:
translations;
resealing operations;
rotation operations;
stretching operations;
trimming operations; and
cutting operations.

10. The method of claim 1 wherein aligning the first and second metric surfaces in three dimensions using the set of one or more surface-modifying operations further comprises:
successively applying surface-modifying operations to one or both of the first and second metric surfaces to minimize the volume between the first and second metric surfaces.

11. An automated system comprising:
one or more processors;
one or more memories;
one or more mass-storage devices; and
computer instructions stored in one or more of the one or more memories that, when retrieved from memory and executed by one or more of the one or more processors, control the automated system to compare a first sequence of time-ordered metric values for a particular metric generated by a first component of a computer system to a second sequence of time-ordered metric values for the particular metric generated by a second component of the computer system to produce a comparison metric that reflects the similarity of the operational behaviors of the different components of the computer system by:
selecting a total time span that includes a sequence of pairs of time-aligned metric values, each pair included a first metric value selected from the first sequence of metric values and a second metric value selected from the second sequence of metric values,
dividing the total time span into multiple time intervals,
for each time interval of the multiple time intervals,
generating, from a portion the first sequence of time-ordered metric values within the time interval, representing time-domain data, a set of frequency-domain data points for the time interval, and
generating, from a portion the second sequence of time-ordered metric values within the time interval, representing two-dimensional time-domain data with time and metric-value axes, a set of frequency-domain data points for the time interval, representing two-dimensional frequency-domain data with frequency and amplitude axes,

stacking the sets of frequency-domain data points generated from portions of the first sequence of time-ordered metric values along a time axis to create a first three-dimensional volume with frequency, time, and amplitude axes,
stacking the sets of frequency-domain data points generated from portions of the second sequence of time-ordered metric values along a time axis to create a second three-dimensional v

Assignees

• Vmware Inc

Inventors

• Knowles Christopher
• Crukley Jeffery

Classifications

• H04L67/131

  Protocols for games, networked simulations or virtual reality · CPC title

• H04L41/142 Primary

  using statistical or mathematical methods · CPC title

• H04L67/1097

  for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS] · CPC title

• H04L67/38

  Electricity · mapped topic