Adaptive Cache Flushing Structure for Memory Efficiency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Modern data storage systems face inefficiencies in memory usage and performance due to excessive memory requirements and contention issues in the shared working-set structure, particularly in scenarios with varying workloads, leading to suboptimal use of leaf structures and increased MID lock contention.

Innovation Solution

The implementation of a dynamically adjustable address range for leaf structures within the shared working-set structure, where the size of address ranges is adjusted based on the PD-to-leaf ratio, allowing for adaptive organization and operation to match workload patterns, thereby improving memory efficiency and reducing contention.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If a fixed-size address range is used for leaf structures in the shared working-set structure, then the system operation is simple and consistent, but memory efficiency deteriorates and MID lock contention increases under varying workloads

Engineering Contradiction:

Improvesystem operation simplicityVSAvoidmemory efficiency

Core Design Contradiction:

Ease of operationVSQuantity of substance

Solution Approach 1:

The patent implements dynamic adjustment of address range sizes for leaf structures based on workload characteristics. The system monitors PD-to-leaf ratios and adjusts the specified size of address ranges between operating periods to optimize memory efficiency while maintaining operational simplicity during each period.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of address range size dynamically. By adjusting the specified size of address ranges based on the PD-to-leaf ratio comparisons, the system adapts to varying workload patterns, improving memory efficiency without compromising the simplicity of operations within each operating period.

Inventive Principle:

Principle #35Parameter changes

2Stability of the object's composition

If a fixed-size address range is used for leaf structures, then the system maintains consistency, but MID lock contention increases under varying workloads

Engineering Contradiction:

Improvesystem consistencyVSAvoidconcurrency performance

Core Design Contradiction:

Stability of the object's compositionVSProductivity

Solution Approach 1:

The system maintains consistency during each operating period while dynamically adjusting address range sizes between periods. This allows the system to adapt to varying workload patterns, reducing MID lock contention and improving concurrency performance without sacrificing operational consistency.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent implements periodic structure-tuning operations between operating periods. During each period, the system operates with a fixed address range size for consistency, then performs adjustments in periodic intervals based on workload analysis, thereby reducing contention while maintaining stability.

Inventive Principle:

Principle #19Periodic action

3Manufacturing precision

If the number of leaf structures is increased to cover more address ranges, then better granularity is achieved, but memory footprint increases

Engineering Contradiction:

Improveaddress range granularityVSAvoidmemory footprint

Core Design Contradiction:

Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent dynamically changes the specified size of address ranges based on the PD-to-leaf ratio. When the ratio indicates underutilization, the system increases address range sizes, reducing the number of leaf structures needed and thereby decreasing memory footprint while maintaining appropriate granularity for the workload.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The system applies different address range sizes locally based on workload characteristics. By adjusting the specified size of address ranges to match actual usage patterns, the system achieves optimal granularity where needed while minimizing memory consumption in areas with lower activity.

Inventive Principle:

Principle #3Local quality

4Shape

If the address range size is decreased for more leaf structures, then better data organization is achieved, but memory efficiency deteriorates

Engineering Contradiction:

Improvedata organization structureVSAvoidmemory efficiency

Core Design Contradiction:

ShapeVSQuantity of substance

Solution Approach 1:

The system dynamically adjusts address range sizes based on workload analysis. By monitoring PD-to-leaf ratios and adapting the specified size of address ranges, the system achieves optimal data organization structure without excessive memory consumption, as the granularity is adjusted to match actual usage patterns.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS11494303B1Data storage system with adaptive, memory-efficient cache flushing structure

Publication Date: 2022.11.08 EMC IP HLDG CO LLC

AI summary

In a method of flushing cached data in a data storage system, instances of a working-set structure (WSS) are used over a succession of operating periods to organize cached data for storing to the persistent storage. In each operating period, leaf structures of the WSS are associated with respective address ranges of a specified size. Between operating periods, a structure-tuning operation is performed to adjust the specified size and thereby dynamically adjust a PD-to-leaf ratio of the WSS, including (1) comparing a last-period PD-to-leaf ratio to a predetermined ratio range, (2) when the ratio is below the predetermined ratio range, increasing the specified size for use in a next operating period, and (3) when ratio is above the predetermined ratio range, then decreasing the specified size for use in the next operating period.