Storage Array Write Aging Model for Cache Destaging Control

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Storage arrays experience performance degradation due to high traffic and repeated write operations, leading to cache memory strain and resource inefficiencies.

Innovation Solution

Implement a multivariate time-series aging model that forecasts optimal write-pending aging times based on local and global I/O characteristics, allowing for dynamic management of write requests in cache memory to extend disk life and optimize resource utilization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If write requests are stored in cache memory for repeated write operations, then I/O performance is improved, but cache memory degrades and system resources are strained

Engineering Contradiction:

ImproveI/O performanceVSAvoidcache memory degradation

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent implements dynamic aging time adjustment based on real-time I/O characteristics. The system continuously monitors local I/O patterns (write-on-write frequency, I/O size) and global I/O characteristics (destage rate, cache size, I/O per second rate) to dynamically modify the aging time for write-pending data, allowing the cache management strategy to adapt to changing workload conditions and prevent cache degradation from repeated writes

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes the aging time parameter based on multiple I/O characteristics. By adjusting this temporal parameter dynamically, the system optimizes when to destage write-pending data from cache to nonvolatile memory, balancing between maintaining I/O performance and preventing cache memory degradation from excessive write operations

Inventive Principle:

Principle #35Parameter changes

2Productivity

If cache memory is used to temporarily store I/O requests, then I/O processing capability is enhanced, but system resources are occupied under significant strain

Engineering Contradiction:

ImproveI/O processing capabilityVSAvoidsystem resources

Core Design Contradiction:

ProductivityVSQuantity of substance

Solution Approach 1:

The system implements selective destaging of write-pending data from cache to nonvolatile memory based on forecasted aging time. By discarding (destaging) data that is unlikely to be rewritten soon and recovering cache space, the system maintains I/O processing capability while reducing the quantity of resources occupied in cache memory under high traffic conditions

Inventive Principle:

Principle #34Discarding and recovering

3Productivity

If storage areas are repeatedly written to in back-to-back operations, then write throughput is maintained, but cache memory is degraded

Engineering Contradiction:

Improvewrite throughputVSAvoidcache memory lifespan

Core Design Contradiction:

ProductivityVSDuration of action of stationary object

Solution Approach 1:

The system uses feedback from monitored I/O characteristics to adjust aging time decisions. By continuously observing local write patterns (write-on-write frequency) and global system state (cache utilization, destage rate), the system provides feedback to the aging model to make informed decisions about when to destage data, preventing excessive cache writes that would degrade cache memory while maintaining write throughput

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS12524179B2Systems and methods for statistical-based IO write pending aging

Publication Date: 2026.01.13 DELL PROD LP

AI summary

Methods and apparatuses for improving storage array performance by using time-series characteristics on both a local storage area (extent-level, or track-level) basis as well as global, system level are disclosed. The time series characteristics are input to an aging model where optimal aging times, on a per-extent basis, are forecasted according to the local and global characteristics. The aging model may be a multivariate time-series model. Incoming write requests may be destaged from cache memory according to the forecasted optimal aging times for the storage extents.