Adaptive Data Packing Scheduling Policy for Cloud Containers

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In cloud computing environments, existing scheduling policies lack dynamic adaptability to workload mixes and resource variability, leading to inefficient resource allocation and increased rejection of large containers due to limited resources and varying resource needs.

Innovation Solution

The implementation of an adaptive data packing scheduling policy that monitors resource variability and workload demand, using a biased sampling algorithm or scheduler plug-in to optimize resource allocation across computing nodes, minimizing rejection probability by equalizing variability in resource needs and availability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If existing scheduling policies are used, then resource allocation is simplified, but resource allocation efficiency deteriorates due to lack of dynamic adaptability

Engineering Contradiction:

Improveresource allocation efficiencyVSAvoidscheduling policy complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The scheduling policy transitions from static to dynamic by continuously monitoring resource variability and workload demand, adjusting container placement decisions in real-time based on current system state rather than following fixed allocation rules

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system implements feedback mechanisms by monitoring resource usage patterns and variability metrics, using this information to inform and adjust subsequent scheduling decisions, creating a closed-loop control system that adapts to changing conditions

Inventive Principle:

Principle #23Feedback

2Reliability

If resource allocation is increased to accommodate all containers, then container rejection is reduced, but resource utilization efficiency deteriorates

Engineering Contradiction:

Improvecontainer acceptance rateVSAvoidresource utilization efficiency

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system changes the parameter of container placement by considering resource variability metrics and workload characteristics, selecting optimal computing nodes based on matched variability profiles rather than simple availability thresholds, thereby improving acceptance rates without excessive resource allocation

Inventive Principle:

Principle #35Parameter changes

3Productivity

If monitoring and adaptive selection are implemented, then resource allocation optimality is improved, but system complexity increases

Engineering Contradiction:

Improveresource allocation optimalityVSAvoidmonitoring and selection system complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The scheduling system performs self-service by automatically monitoring its own resource state and making adaptive placement decisions without external intervention, using built-in variability metrics and workload analysis to optimize container allocation autonomously

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS10887246B2Adaptive data packing

Publication Date: 2021.01.05 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

Techniques and a system for adaptive data packing are provided. In one example, a system includes a container component and a computing node component. The container component monitors one or more resources for container data in a network environment and determines variability data representative of a variability of the one or more resources for the container data during a period of time. The computing node component that selects a computing node in the network environment for particular container data based on the variability data.