Assessing performance of networked computing environments

Patent metadata

Field	Value
Publication number	US-10601690-B2
Application number	US-201916277792-A
Country	US
Kind code	B2
Filing date	Feb 15, 2019
Priority date	Mar 27, 2015
Publication date	Mar 24, 2020
Grant date	Mar 24, 2020

Abstract

Systems and methods are described for accurately determining which of a set of networked computing environments most closely approximates a target computing environment. The determination is based on executing a reference workload in the target computing environment and calculating a coefficient of equivalency for each of the networked computing environments, which relates the performance of one or more computing resources in the target computing environment to the corresponding resources in the networked computing environments. The coefficient of equivalency may further be used to determine which of a set of networked computing environments will provide a target level of performance when executing the workload. The target computing environment may be assessed in terms of time, cost, percentage of capacity utilized, or other criteria.

First claim

What is claimed is:

1. A computer-implemented method comprising:
causing execution of a workload in a networked computing environment, the networked computing environment providing one or more networked computing resources associated with a target computing environment comprising one or more target computing resources with specified resource metrics in accordance with the execution of the workload;
monitoring utilization of the one or more networked computing resources, wherein monitoring the utilization for each of the one or more networked computing resources comprises collecting a utilization metric;
determining a resource profile for the networked computing environment associated with an aggregate utilization of the one or more networked computing resources of the networked computing environment;
modeling, in the networked computing environment, execution of the workload in the target computing environment;
determining a resource profile for the target computing environment associated with an aggregate utilization for the target computing environment; and
generating, for individual networked computing environments, a first expression representative of a comparison of a first total utilization of the one or more networked computing resources of the networked computing environment to a first total utilization of the target computing environment, wherein the first expression is based at least in part on the workload, the resource profile for the first networked computing environment, and the resource profile for the target computing environment.

2. The computer-implemented method of claim 1 , wherein the resource profile for the networked computing environment includes at least one performance metric for at least one networked computing resource, wherein individual networked computing resources correspond to a target computing resource.

3. The computer-implemented method of claim 1 , wherein the first total utilization for the target computing environment is based at least in part on one of a total duration of utilization for at least one target computing resource, a total cost of utilization for at least one target computing resource, or a total percentage of capacity utilized for at least one target computing resource.

4. The computer-implemented method of claim 1 , wherein modeling execution of the workload comprises:
calculating a resource coefficient for individual target computing resources, the resource coefficient based at least in part on the specified resource metric for the individual target computing resource and the specified resource metric for a corresponding networked computing resource;
predicting a utilization metric for the individual target computing resources, the predicted utilization metric for the individual target computing resource based at least in part on the resource coefficient for the individual target computing resource and a utilization metric for the corresponding networked computing resource; and
predicting an unavailability metric for the individual target computing resources, the predicted unavailability metric for the individual target computing resource based at least in part on the resource coefficient for the individual target computing resource and an unavailability metric for the corresponding networked computing resource.

5. The computer-implemented method of claim 1 further comprising:
obtaining a second resource profile for the target computing environment, the second resource profile for the target computing environment including at least a second total utilization of the target computing environment when a second workload is executed in the target computing environment;
generating, for individual networked computing environments, a second expression that represents a second total utilization of the one or more networked computing resources of the networked computing environment relative to the second total utilization of the target computing environment, wherein the second expression is based at least in part on the second resource profile for the target computing environment and the resource profile for the individual networked computing environment; and
selecting, based at least in part on the second expression for the individual networked computing environments, a second networked computing environment from the plurality of networked computing environments that approximates the target computing environment.

6. The computer-implemented method of claim 5 further comprising selecting, based at least in part on the first expression for the individual networked computing environments and the second expression for the individual networked computing environments, a third networked computing environment from the plurality of networked computing environments that approximates the target computing environment.

7. The computer-implemented method of claim 1 , wherein the first expression is a numeric expression.

8. The computer-implemented method of claim 7 , wherein the numeric expression is selected from a set of defined values.

9. A system comprising:
a non-transitory computer-readable data store configured to store specific computer-executable instructions; and
a computing device including a processor in communication with the non-transitory computer-readable data store, the processor configured to execute the computer-executable instructions to:
obtain one or more resource metrics for individual target computing resources in a target computing environment;
cause execution of a workload in a networked computing environment, the networked computing environment providing one or more networked computing resources associated with the individual target computing resources in the target computing environment and associated with specified resource metrics;
collect a resource metric and calculate a utilization metric for each of the one or more networked computing resources;
determine a resource profile for the networked computing environment associated with an aggregate duration for the networked computing environment based at least in part on durations of calculated utilization metrics for the individual networked computing resources;
determine a resource profile for the target computing environment associated with an aggregate duration for the target computing environment based at least in part on the utilization metrics for the individual target computing resources; and
generate a first expression based at least in part on the workload, the resource profile for the networked computing environment, and the resource profile for the target computing environment.

10. The system of claim 9 , wherein the individual target computing resources comprise at least one of a central processing unit, a memory, a network interface, a computer-readable storage medium, a graphics processing unit, or an operating system.

11. The system of claim 9 , the processor further configured to execute the computer-readable instructions to model execution of the workload in the target computing environment, wherein modeling the execution of the workload in the target computing environment comprises predicting a duration of utilization for the individual target computing resources, the prediction based at least in part on the resource metric for the target computing resource, the resource metric for a corresponding networked computing resource, and the utilization metric for the corresponding networked resource.

12. The system of claim 11 , wherein calculating the utilization metric of the one or more networked computing resources further comprises calculating a duration of unavailability for the individual networked computing resources, and wherein the

Assignees

• Amazon Tech Inc

Inventors

• Sait Abdul Sathar

Classifications

• G06F11/3006

  where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems (multiprogramming arrangements G06F9/46; allocation of resources G06F9/50) · CPC title

• H04L43/0876 Primary

  Network utilisation, e.g. volume of load or congestion level · CPC title

• G06F2009/45591

  Monitoring or debugging support · CPC title

• G06F11/3414

  Workload generation, e.g. scripts, playback · CPC title

• G06F11/3433

  for load management (allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029) · CPC title