Virtual Block Temperature Metrics for Storage Tier Placement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data storage systems face inefficiencies in tracking and updating metrics for content placement across different storage tiers, particularly in multi-tiered systems where SSDs and HDDs are used, as they lack effective mechanisms to determine the optimal placement of data based on activity levels and access frequencies.
Innovation Solution
Implementing a method where multiple processing cores record read misses and overwrites in tablets, switch roles based on criteria, and merge entries to update metric sets, which are then used to determine temperature classifications for content placement across storage tiers, allowing for efficient promotion and demotion of data between SSD and HDD tiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data storage systems track and update metrics for content placement across different storage tiers, then storage efficiency and performance are improved, but system complexity increases due to the need for monitoring, recording, and processing activity metrics
Solution Approach 1:
The system divides the storage system into multiple storage tiers (e.g., SSD tier, HDD tier) and segments the activity metrics tracking by creating separate data structures for different storage entities. Each storage tier has its own metrics tracking mechanism, allowing independent optimization without affecting the entire system.
Solution Approach 2:
The patent introduces an intermediary layer of activity metrics data structures that mediate between the storage controllers and the placement decisions. These intermediaries collect and process activity information without requiring direct complex interactions between all system components, simplifying the overall architecture.
2Measurement precision
If the system continuously monitors and updates activity metrics for all storage entities, then placement accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The system performs preliminary actions by pre-defining the structure of activity metrics data structures and pre-establishing the mapping between storage entities and their metrics. This preparation allows for faster runtime processing and reduces the computational overhead during continuous monitoring.
Solution Approach 2:
The patent dynamically adjusts monitoring parameters based on storage entity types and access patterns. Different storage entities (e.g., files, volumes, LUNs) have different metric collection frequencies and granularities, allowing the system to optimize processing time while maintaining placement accuracy through parameter adaptation.
Data Source
AI summary
In a multi-tiered system, read temperature and write temperature per virtual block (VLB) can be determined based on the content referenced by each VLB. The read temperature can describe VLB read activity relative to the recency and frequency of read misses of content referenced by the VLB. The write temperature can describe VLB write activity relative to the recency and frequency of overwrites with respect to content referenced by the VLB. Read misses and overwrites can be recorded per processing core in an active tablet used for recording by only a single core. Active and analytic tablet sets across cores can be switched based on criteria. Analytic tablets can be processed and used to update metrics, including read and write temperatures, stored for VLBs. The metrics can be used to determine temperature classifications of VLBs. Content referenced by VLBs can be promoted and/or demoted using VLB temperature classifications.