Composable Infrastructure Resource Allocation via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing infrastructure often experiences underutilization of resources, with CPU utilization at most 50% and peripheral infrastructure utilization below 70%, leading to significant financial losses due to underutilized resources.
Innovation Solution
A system that employs machine learning and artificial intelligence to dynamically allocate and manage resources within a composable infrastructure, optimizing workload distribution across available hardware resources by predicting and implementing actions that improve resource utilization and meet service level requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If cloud computing infrastructure is deployed to provide computing resources, then service capacity and availability are improved, but resource utilization deteriorates (CPU utilization at most 50%, peripheral infrastructure utilization below 70%)
Solution Approach 1:
The patent implements dynamic resource allocation by continuously monitoring workload demands and automatically adjusting resource provisioning in real-time. The system transitions from static infrastructure to dynamic resource management, where computing, storage, and networking resources are flexibly allocated based on actual needs, enabling both high service capacity and improved utilization.
Solution Approach 2:
The patent creates a universal resource pool where infrastructure components can serve multiple workloads and functions simultaneously. By abstracting and virtualizing underlying resources, the system enables single resources to fulfill multiple roles across different applications and services, thereby increasing overall utilization while maintaining service capacity.
2Reliability
If infrastructure resources are increased to meet growing workload demands, then service level requirements are improved, but financial loss increases due to underutilized resources
Solution Approach 1:
The patent implements closed-loop feedback mechanisms that continuously monitor service level agreement compliance, workload performance, and resource utilization metrics. The system uses this feedback to automatically adjust resource allocation, scaling resources up when service levels are at risk and scaling down when utilization is low, thereby maintaining reliability while minimizing financial loss from underutilization.
Solution Approach 2:
The patent dynamically changes key operational parameters such as resource provisioning levels, workload distribution patterns, and infrastructure configuration based on real-time conditions. By adjusting these parameters in response to changing demands and service level requirements, the system optimizes the balance between reliability and cost efficiency.
3Device complexity
If static resource allocation is used to simplify infrastructure management, then device complexity is reduced, but resource utilization deteriorates
Solution Approach 1:
The patent implements self-service automation where the infrastructure management system autonomously performs resource allocation, workload placement, and infrastructure configuration without requiring complex manual intervention. The system automatically discovers workload requirements, selects appropriate resources, and provisions infrastructure components, simplifying management while achieving optimal utilization through intelligent automation.
Data Source
AI summary
Apparatuses, systems, and techniques to select action(s) predicted to modify at least one current state of a computing system using values of at least one parameter, values of at least one system objective, and at least one desired state of the computing system defined at least in part by the at least one system objective, and provide the action(s) to an application that implements the action(s) with respect to the computing system.