Tuning a container orchestrator
AssigneeInternational Business Machines Corporation
InventorGireesh Punathil
Life of this patent
Estimating the term…
Abstract
A method includes determining, by a computing device, a programming language of a running workload; selecting, by the computing device, a profiler module in view of the programming language; determining, by the computing device, characteristics of the running workload using the profiler module; identifying, by the computing device, a persona of the running workload from the characteristics; identifying, by the computing device, orchestration policies which map to the persona; tuning, by the computing device, a container orchestrator module in view of the orchestration policies; and deploying, by the computing device, containers to the running workload using the tuned container orchestrator module.
Claims20
A method, comprising:
- determining, by a computing device, a programming language of a running workload;
- selecting, by the computing device, a profiler module in view of the programming language;
- determining, by the computing device, characteristics of the running workload using the profiler module;
- identifying, by the computing device, a persona of the running workload from the determined characteristics by utilizing machine learning to train identity data of the persona and learn the persona which corresponds with the determined characteristics;
- identifying, by the computing device, orchestration policies which map to the persona by utilizing machine learning to learn which of the orchestration policies map to the persona;
- tuning, by the computing device, a container orchestrator module in view of the orchestration policies;
- and
- deploying, by the computing device, containers to the running workload using the tuned container orchestrator module,
- wherein the determined characteristics include a lower input of homogenous data, a high output of homogenous data, and a fluctuating memory profile of a streaming workload.
A computer program product comprising one or more computer readable storage media having program instructions collectively stored on the one or more computer readable storage media, the program instructions executable to:
- determine a workload is running;
- determine a programming language of the workload;
- select a profiler module in view of the programming language;
- determine characteristics of the workload using the profiler module;
- identify a persona of the workload from the determined characteristics by utilizing machine learning to train identity data of the persona and learn the persona which corresponds with the determined characteristics;
- identify orchestration policies which map to the persona by utilizing machine learning to learn which of the orchestration policies map to the persona;
- and
- tune a container orchestrator module in view of the orchestration policies,
- wherein the determined characteristics include a lower input of homogenous data, a high output of homogenous data, and a fluctuating memory profile of a streaming workload.
A system comprising:
- a processor, a computer readable memory, one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable to:
- determine a container orchestrator module is performing actions to run a workload;
- determine a programming language of the workload;
- select a profiler module in view of the programming language;
- determine characteristics of the workload using the profiler module;
- identify a persona of the workload from the determined characteristics by utilizing machine learning to train identity data of the persona and learn the persona which corresponds with the determined characteristics;
- identify orchestration policies which map to the persona by utilizing machine learning to learn which of the orchestration policies map to the persona;
- tune the container orchestrator module in view of the orchestration policies;
- and
- deploy containers to the workload using the tuned container orchestrator module,
- wherein the determined characteristics include a lower input of homogenous data, a high output of homogenous data, and a fluctuating memory profile of a streaming workload.
Dependencies are inferred from the claim text.