Idle Compute Resource Scheduling for Parallel Machine Learning Tasks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems and techniques for performing machine learning (ML) and artificial intelligence (AI) applications lack sufficient processing power, leading to inefficiencies, extended task performance times, and increased reliance on external cloud platforms, resulting in productivity delays and infrastructure costs.
Innovation Solution
Identify and leverage unused computing resources within an organization's internal cloud environment to execute ML tasks by breaking them into self-contained sub-tasks, de-identifying data, and reassembling the results, while utilizing both internal and external cloud resources as needed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional systems sequentially perform ML/AI tasks on dedicated computing resources, then task completion is guaranteed, but processing time is excessively long and computing resources are occupied for extended periods
Solution Approach 1:
The patent segments ML/AI tasks into smaller sub-tasks that can be executed independently on multiple computing nodes simultaneously. This allows parallel processing of task components, significantly reducing overall execution time while maintaining task completion integrity.
Solution Approach 2:
The patent enables computing resources to serve multiple functions by allowing idle nodes to process ML tasks while active nodes continue their primary workloads. This multi-functionality increases productivity without requiring dedicated exclusive resources for ML tasks.
2Power
If more computing resources are allocated to ML/AI tasks, then processing power increases, but infrastructure costs increase
Solution Approach 1:
The patent implements a self-service model where the system automatically identifies and utilizes idle computing resources within the organization. This eliminates the need to purchase additional dedicated infrastructure for ML tasks, as existing resources serve themselves by processing tasks during idle periods.
Solution Approach 2:
Existing computing infrastructure performs dual functions: primary workloads during active periods and ML task processing during idle periods. This universality increases effective processing power without requiring additional infrastructure investment.
3Reliability
If dedicated computing infrastructure is used for ML tasks, then task performance is reliable, but reliance on external cloud platforms increases
Solution Approach 1:
The patent extracts ML task processing capability from external cloud platforms and implements it within the organization's own computing infrastructure. This reduces external dependency while maintaining reliable execution through internal resource management and scheduling.
Data Source
AI summary
Systems and methods relating to leveraging inactive computing resources are discussed. An example system may include one or more computing nodes having an active state and an inactive state, one or more processors, and a memory. The memory may contain instructions therein that, when executed, cause the one or more processors to identify a task to be performed by the one or more computing nodes based upon a received request. The instructions may further cause the one or more processors to create one or more sub-tasks based upon the task and schedule the one or more sub-tasks for execution on the one or more computing nodes during the inactive state. The instructions may further cause the one or more processors to collate the one or more sub-tasks into a completed task, and generate a completed task notification based upon the completed task.