Cloud Compute Clearing System for Dynamic Job Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing systems face inefficiencies due to over-provisioning of resources and the complexity of identifying job constraints, leading to increased costs and resource waste, as consumers struggle to optimize compute tasks across multiple providers.
Innovation Solution
A cloud compute clearing system that automatically identifies job constraints and optimizes resource allocation by partitioning tasks across multiple providers based on resource availability and translation costs, ensuring timely and budget-friendly execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If consumers manually identify and optimize job constraints across multiple compute providers, then job execution efficiency can be improved, but the complexity and time required for information gathering and decision-making increases significantly
Solution Approach 1:
The patent introduces a cloud compute clearing system as an intermediary between consumers and multiple compute providers. This clearing system automatically identifies job constraints, partitions tasks, and optimizes resource allocation across providers, eliminating the need for consumers to manually gather information and make complex decisions about constraint identification and provider selection.
Solution Approach 2:
The clearing system performs self-service by automatically analyzing job requirements, identifying appropriate constraints, and making optimization decisions without requiring consumer intervention. The system autonomously manages the complex task of matching jobs to suitable compute providers based on multiple factors including resource availability and pricing.
2Ease of operation
If consumers directly manage task allocation across multiple compute providers, then control over job execution is maximized, but the time and expertise required for optimization increases
Solution Approach 1:
The patent extracts the complex optimization function from the consumer's responsibility and transfers it to the clearing system. Consumers simply submit their compute tasks with basic requirements, while the clearing system extracts and processes the detailed constraint identification and provider selection tasks, significantly reducing the time and expertise burden on consumers.
3Reliability
If compute providers over-provision physical resources to ensure sufficient compute capacity, then service reliability is improved, but resource waste and operational costs increase
Solution Approach 1:
The clearing system implements dynamic resource allocation by continuously monitoring compute provider capacity and pricing, and adjusting task distribution in real-time. This dynamic approach allows providers to serve more variable workloads without permanent over-provisioning, improving resource utilization while maintaining reliability through on-demand capacity allocation.
Solution Approach 2:
The system changes the parameter of resource allocation from static over-provisioning to dynamic optimization based on multiple factors including current capacity availability, pricing conditions, and job requirements. This allows the system to achieve reliable service delivery without the constant resource overhead of traditional over-provisioning strategies.
4Adaptability or versatility
If the cloud compute market differentiates based on multiple factors such as price, reliability, and manageability, then service quality improves, but the complexity of identifying and comparing underlying capacities overwhelms consumers
Solution Approach 1:
The clearing system provides a universal interface that handles multiple differentiation factors (price, reliability, manageability) through a single automated optimization process. Instead of requiring consumers to evaluate and compare multiple providers across numerous dimensions, the clearing system universally applies multi-criteria optimization to select the most appropriate provider for each task based on all relevant factors.
Data Source
AI summary
Provided are systems and methods for simplifying cloud compute markets. A compute marketplace can be configured to determine, automatically, attributes and/or constraints associated with a job without requiring the consumer to provide them. The compute marketplace provides a clearing house for excess compute resources which can be offered privately or publically. The compute environment can be further configured to optimize job completion across multiple providers with different execution formats, and can also factor operating expense of the compute environment into the optimization. The compute marketplace can also be configured to monitor jobs and/or individual job partitions while their execution is in progress. The compute marketplace can be configured to dynamically redistribute jobs/job partitions across providers when, for example, cycle pricing changes during execution, providers fail to meet defined constraints, excess capacity becomes available, compute capacity becomes unavailable, among other options.