PBS Job Conversion to Kubernetes Configurations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The use of Portable Batch System (PBS) in computer clusters is costly and inefficient, with limited resources leading to delayed job execution and wasted power consumption due to underutilization of resources.

Innovation Solution

A system and method for converting PBS job inputs into Kubernetes job configuration files, allowing PBS jobs to be executed in a Kubernetes environment, thereby improving resource utilization and efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If PBS is used for cluster management, then job execution is possible, but resource utilization is low and power consumption is wasted

Engineering Contradiction:

Improveresource utilizationVSAvoidpower consumption

Core Design Contradiction:

ProductivityVSLoss of energy

Solution Approach 1:

The patent introduces a conversion tool as an intermediary that translates PBS job scripts into Kubernetes job configuration files. This mediator enables seamless migration from PBS to Kubernetes without requiring users to learn new tools, while leveraging Kubernetes' superior resource management capabilities to improve resource utilization and reduce energy waste

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If PBS cluster resources are increased, then more jobs can be handled, but cost increases

Engineering Contradiction:

Improvejob handling capacityVSAvoidcluster resources

Core Design Contradiction:

ProductivityVSQuantity of substance

Solution Approach 1:

The patent makes the cluster management system universal by adopting Kubernetes, which can handle diverse workloads including batch jobs, containers, and orchestration tasks. This multi-functional platform replaces the specialized PBS system, allowing the same cluster resources to serve multiple purposes and increasing overall job handling capacity without proportional resource increases

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Loss of time

If PBS cluster size is limited, then cost is controlled, but job execution is delayed

Engineering Contradiction:

Improvejob execution delayVSAvoidjob throughput

Core Design Contradiction:

Loss of timeVSProductivity

Solution Approach 1:

The patent leverages Kubernetes' dynamic resource allocation and scheduling capabilities, which can automatically adjust resource distribution based on current workload demands. This dynamic approach eliminates job execution delays by efficiently managing limited cluster resources, ensuring high job throughput without increasing cluster size

Inventive Principle:

Principle #15Dynamics

4Reliability

If PBS operators are hired, then cluster maintenance is possible, but operational cost increases

Engineering Contradiction:

Improvecluster maintenanceVSAvoidoperational cost

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The patent adopts Kubernetes' automated self-service capabilities, including self-healing, automatic scaling, and built-in monitoring. These features enable the cluster to maintain itself without requiring human operators, reducing operational costs while ensuring reliability through automated maintenance tasks

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS11487555B2Running PBS jobs in kubernetes

Publication Date: 2022.11.01 TENCENT AMERICA LLC

AI summary

A system is configured to run jobs in Kubernetes based on PBS job inputs. The system may convert a PBS input, such as in the form of a PBS job command line input and/or a PBS job script, to a Kubernetes job configuration file. A Kubernetes system may then perform a job according to the Kubernetes job configuration file. Through implementation of the conversion process, PBS users can leverage the capabilities of Kubernetes to have jobs performed without having to know how to use Kubernetes.