Workload Resource Estimation for Cloud Bursting

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current cloud bursting mechanisms rely heavily on manual monitoring and resource management, leading to inefficiencies in identifying resource shortages or excesses, which can result in overprovisioning or underprovisioning of resources, increasing costs and potential downtime.

Innovation Solution

A computer program and method that calculates the required resources for unfinished workloads by considering average completion durations and consumed resources across workload classes, adjusting for remaining processing time and provisioning duration, to accurately determine the necessary resources needed for workload processing and optimize resource allocation between local and cloud systems.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual monitoring and resource management are used in cloud bursting mechanisms, then operational flexibility is maintained, but resource provisioning accuracy deteriorates leading to overprovisioning or underprovisioning

Engineering Contradiction:

Improveresource provisioning accuracyVSAvoidresource allocation inefficiency

Core Design Contradiction:

Measurement precisionVSLoss of energy

Solution Approach 1:

The system continuously monitors workload conditions and uses this feedback to dynamically adjust resource provisioning decisions. The workload manager receives real-time information about workload status, resource utilization, and completion durations, then automatically adjusts cloud host provisioning accordingly, eliminating the need for manual monitoring while maintaining high accuracy.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The cloud bursting mechanism operates autonomously with the workload manager automatically detecting resource needs and provisioning cloud hosts without human intervention. The system self-manages the entire resource allocation process by calculating required resources based on workload characteristics and historical data, then automatically requesting and releasing cloud hosts as needed.

Inventive Principle:

Principle #25Self-service

2Reliability

If cloud hosts are provisioned to accommodate peak resource demand, then resource availability is improved, but total infrastructure cost increases due to underutilization during low-demand periods

Engineering Contradiction:

Improveresource availabilityVSAvoidtotal infrastructure cost

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The system dynamically adjusts the number of provisioned cloud hosts based on real-time workload conditions rather than maintaining a static infrastructure. The workload manager continuously evaluates whether cloud hosts are needed and adjusts provisioning accordingly, ensuring resource availability when required while minimizing infrastructure costs during low-demand periods.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes the provisioning parameter (number of cloud hosts) based on workload characteristics such as completion duration, resource consumption patterns, and current queue status. By adjusting this parameter dynamically according to actual needs rather than peak demand, the system maintains reliability while optimizing cost.

Inventive Principle:

Principle #35Parameter changes

3Adaptability or versatility

If cloud bursting mechanisms are implemented, then scalability is improved, but complexity of resource management increases

Engineering Contradiction:

ImprovescalabilityVSAvoidresource management complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The workload manager performs multiple functions including monitoring workload status, calculating resource requirements, making provisioning decisions, and managing cloud host lifecycle. By consolidating these diverse functions into a single multi-functional component, the system achieves scalability without proportionally increasing management complexity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Loss of energy

If excess cloud host capacity is returned to cloud providers, then cost savings are achieved, but risk of resource shortage during workload spikes increases

Engineering Contradiction:

Improvecost efficiencyVSAvoidresource sufficiency

Core Design Contradiction:

Loss of energyVSReliability

Solution Approach 1:

The system maintains a queue of pending workloads and uses this information to predict future resource needs. By looking ahead at the workload pipeline and calculating expected completion durations, the system proactively provisions resources before they are needed while still returning excess capacity to minimize costs. This preliminary action based on workload forecasting balances cost efficiency with resource sufficiency.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11573835B2Estimating resource requests for workloads to offload to host systems in a computing environment

Publication Date: 2023.02.07 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

Provided are a computer program product, system, and method for estimating resource requests for workloads to offload to host systems in a computing environment. A calculation is made required resources of computational resources required to complete processing a plurality of unfinished workloads that have not completed. A determination is made of allocated resources that are not yet provisioned to workloads. The required resources are reduced by the allocated resources not yet provisioned to determine resources to provision. The resources to provision for the unfinished workloads are requested.