Logical Cluster Partitioning for Dynamic Job Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current job scheduling techniques in networked computer clusters require users to explicitly specify node types and cannot be configured dynamically, failing to consider workload attributes, leading to inefficient task execution.
Innovation Solution
A system that dynamically determines sets of homogenous nodes by evaluating node characteristics and attributes, allowing for dynamic scheduling based on task requirements, using a scheduler with components like logical partition filters and evaluators to identify and rank suitable node partitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If users explicitly specify node types during deployment to generate node partitions, then task scheduling can be performed with known node characteristics, but the system cannot be configured dynamically and cannot consider workload attributes
Solution Approach 1:
The patent implements dynamic node partitioning where the scheduler evaluates node attributes and workload requirements in real-time to create partitions on-the-fly, rather than using static pre-defined partitions. This allows the system to adapt to different workload characteristics dynamically while maintaining manageable complexity through automated evaluation algorithms.
Solution Approach 2:
The system changes the parameters used for partitioning from fixed node types specified during deployment to dynamic attributes evaluated at scheduling time. The scheduler considers multiple node attributes (hardware specifications, current load, availability) and workload characteristics to determine optimal partitions, allowing flexible adaptation without requiring complex manual reconfiguration.
2Productivity
If static node partitions are used without considering workload attributes, then the scheduling system is simpler to implement, but resource utilization efficiency decreases
Solution Approach 1:
The system performs preliminary evaluation of node attributes and creates candidate partitions before actual task assignment. The scheduler pre-assesses node characteristics and prepares multiple partition options that can be quickly selected based on workload requirements, improving resource utilization without requiring complex real-time decisions during task execution.
Solution Approach 2:
The scheduling system incorporates feedback mechanisms where the evaluation of node partitions considers both static node attributes and dynamic workload characteristics. The system continuously monitors and adjusts partition selections based on current system state and task requirements, optimizing resource utilization through informed decision-making while maintaining manageable complexity through structured evaluation criteria.
Data Source
AI summary
Apparatuses, systems, and techniques to select computer systems to perform portions of one or more programs in parallel based, at least in part, on the computer systems' ability to perform the portions at substantially a same performance level. In at least one embodiment, a system includes one or more circuits to select one or more computer systems based, at least in part, on identifying one or more logical partitions of the computer systems based, at least in part, on one or more attributes of one or more programs associated with the one or more computer systems.