Adaptive IO Pattern Detection for Data Pre-fetching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Data protection systems face performance issues due to non-sequential access patterns, making it difficult to predict future data requests and resulting in lower cache hit ratios and reduced system performance.

Innovation Solution

Implementing a system and method to detect and analyze Input/Output (IO) patterns in data protection systems, allowing for pre-fetching of data into cache based on access patterns, even when they are not strictly sequential or random, by tracking accesses at a storage system level and adapting pre-fetching strategies as patterns change over time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If data is pre-fetched based on sequential access patterns, then cache hit ratio improves, but system adaptability deteriorates when access patterns become non-sequential

Engineering Contradiction:

Improvecache hit ratioVSAvoidadaptability to non-sequential access patterns

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The system dynamically adapts its pre-fetching strategy based on detected access patterns. It transitions from static sequential pre-fetching to dynamic pattern-based pre-fetching that can handle sequential, random, and streaming access patterns, thereby maintaining high cache hit ratios across varying workloads

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system implements feedback mechanisms by monitoring actual access patterns and using this information to adjust pre-fetching decisions. Access pattern detection provides feedback that modifies future pre-fetching behavior, enabling the system to adapt to changing access characteristics and maintain optimal performance

Inventive Principle:

Principle #23Feedback

2Device complexity

If pre-fetching strategy is simplified to handle sequential patterns, then implementation complexity reduces, but measurement precision of access patterns deteriorates

Engineering Contradiction:

Improvepre-fetching implementation complexityVSAvoidaccess pattern detection accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The access pattern detection is segmented into distinct pattern types (sequential, random, streaming) with specific detection criteria for each. This segmentation allows the system to measure and identify different access patterns accurately using targeted detection mechanisms, improving overall measurement precision without requiring a monolithic complex system

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system changes detection parameters and thresholds based on the specific access pattern being measured. Different access patterns trigger different detection parameters, allowing precise measurement of each pattern type while managing implementation complexity through parameterized detection rather than monolithic analysis

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11567871B2Input/output patterns and data pre-fetch

Publication Date: 2023.01.31 EMC IP HLDG CO LLC

AI summary

Systems and methods for determining an access pattern in a computing system. Accesses to a file may contain random accesses and sequential accesses. The file may be divided into multiple regions and the accesses to each region are tracked. The access pattern for each region can then be determined independently of the access patterns of other regions of the file.