Multi-Tier Storage Deployment for Changing Workloads and Device Costs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-tier storage systems face challenges in optimizing cost, performance, and endurance due to static upfront workload analysis, which fails to account for changing device selection and price over time, leading to suboptimal device configuration.
Innovation Solution
A storage management engine determines the optimal deployment configuration for multi-tier storage systems by analyzing available persistent memory devices, their characteristics, and user requirements, providing dynamic recommendations for device selection and configuration adjustments based on runtime data to minimize storage acquisition cost and total cost of ownership.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If static upfront workload analysis is used for storage system configuration, then initial deployment is simplified, but the system cannot adapt to changing workloads and device price fluctuations over time
Solution Approach 1:
The storage management system transitions from static upfront configuration to dynamic runtime optimization. The engine continuously monitors workload characteristics, device performance, and pricing information, automatically adjusting data placement and migration strategies to adapt to changing conditions without requiring complex manual reconfiguration.
Solution Approach 2:
The system implements self-service through automated workload analysis and device selection. The storage management engine independently evaluates workload requirements, compares available storage devices against multiple criteria including performance, cost, and endurance, and makes autonomous decisions about data placement and migration without requiring continuous human intervention.
2Quantity of substance
If device selection and configuration are optimized dynamically, then storage acquisition cost is reduced, but real-time monitoring and analysis infrastructure is required
Solution Approach 1:
The storage management engine performs multiple functions within a single integrated system: workload characterization, device evaluation, cost analysis, and automated data migration. This multi-functional approach consolidates what would otherwise require separate monitoring, analysis, and management systems into one unified platform, reducing overall infrastructure complexity while enabling dynamic optimization.
Solution Approach 2:
The system implements continuous feedback loops where the management engine monitors actual workload patterns, device performance metrics, and cost parameters, then uses this information to iteratively improve data placement decisions. Runtime workload samples are fed back into the optimization engine, which adjusts migration strategies based on observed patterns, enabling progressive cost reduction without requiring overly complex external monitoring infrastructure.
3Productivity
If data is stored in multiple storage tiers, then performance and cost are optimized, but data migration between tiers increases system complexity
Solution Approach 1:
The system manages multi-tier storage by dynamically changing parameters such as data hotness classification, access frequency thresholds, and device selection criteria. The management engine adjusts these parameters based on runtime workload analysis, automatically determining when and how to migrate data between storage tiers. This parameter-driven approach simplifies migration complexity compared to fixed threshold methods, as the system adapts migration triggers to actual usage patterns rather than requiring complex predetermined rules.
Data Source
AI summary
An embodiment of an electronic apparatus may comprise a processor, memory communicatively coupled to the processor, and circuitry communicatively coupled to the processor and the memory to determine a group of available types of persistent memory devices and a set of characteristics associated with each type of persistent memory device of the group of available types of persistent memory devices, determine of a set of requirements for a storage system, and determine a deployment configuration for the storage system with a lowest storage acquisition cost based on the group of available types of persistent memory devices, the sets of characteristics, and the set of requirements. Other embodiments are disclosed and claimed.