Tiered Memory Fabric Workload Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional workload performance optimization techniques are inadequate for systems with multiple heterogeneous processing systems connected to a pool of memory systems, as they fail to account for varying performance requirements and utilization levels across different processing and memory systems.
Innovation Solution
A workload management system that identifies and configures respective fabric processing systems and memory systems within a tiered memory fabric to optimize workload performance pipelines by allocating memory systems based on performance characteristics, ensuring efficient resource utilization and satisfying Service Level Agreements (SLAs).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional workload performance optimization techniques are used with static connectivity, then system simplicity is maintained, but workload performance cannot be optimized across heterogeneous processing systems
Solution Approach 1:
The patent implements dynamic workload routing that adapts to varying workload requirements and system states. The workload management engine continuously monitors performance metrics and dynamically assigns workloads to processing systems and memory systems based on current conditions, transforming the static connectivity model into a dynamic optimization system that resolves the contradiction between performance improvement and system complexity
Solution Approach 2:
The system changes multiple parameters simultaneously including workload assignment, memory system selection, and connectivity configuration. By dynamically adjusting these parameters based on workload characteristics and system state, the patent achieves performance optimization across heterogeneous systems while managing complexity through coordinated parameter transformation
2Productivity
If higher-performance memory systems are allocated to all processing systems, then workload performance is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent applies local quality by allocating memory systems with specific performance characteristics to processing systems based on their individual workload requirements. Instead of uniformly allocating high-performance memory to all systems, the workload management engine matches memory allocation to local needs, ensuring high-performance memory is used only where necessary while maintaining resource utilization efficiency
Solution Approach 2:
The system dynamically changes memory allocation parameters based on workload characteristics and system state. The workload management engine adjusts which memory systems are allocated to which processing systems in real-time, transforming the resource allocation strategy to balance performance requirements with resource utilization efficiency
3Productivity
If memory systems are allocated without considering workload requirements, then allocation simplicity is maintained, but workload performance optimization is lost
Solution Approach 1:
The patent implements preliminary action by having the workload management engine proactively analyze workload requirements and pre-determine optimal memory system allocations before workloads are executed. This advance planning enables performance optimization while managing allocation complexity through systematic preliminary assessment and configuration
Solution Approach 2:
The system employs feedback mechanisms where the workload management engine monitors workload performance and memory system utilization, using this information to continuously optimize memory allocations. The feedback loop transforms complex allocation decisions into manageable iterative improvements based on actual system performance data
Data Source
AI summary
A tiered memory fabric workload performance optimization system includes a workload management device coupled to a processing fabric and a memory fabric. The workload management system receives a workload request to perform a workload including sub-workloads, and identifies a respective processing system in the processing fabric for performing each of the sub-workloads. The workload management device then determines, for use by each respective processing system identified for performing the sub-workloads, a respective memory system in the memory fabric to provide memory systems in different memory tiers in the memory fabric that optimize characteristic(s) of a workload performance pipeline provided by the respective processing systems identified for performing the sub-workloads. The workload management device then configures each respective processing system identified for performing each of the sub-workloads, and the respective first memory system determined for that respective processing system, to perform the sub-workload that respective processing system was identified to perform.