Ghost List Cache Eviction with Recency and Frequency Adjustment

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing cache eviction techniques, such as time threshold-based methods, often evict frequently used entries due to fixed thresholds, leading to suboptimal cache performance as less frequently used entries remain in the cache, thereby affecting overall system efficiency.

Innovation Solution

Implementing an adjustment mechanism that adjusts timestamps of cache entries based on recency and frequency, using ghost lists to extend the time to live (TTL) for frequently used entries, thereby improving cache eviction strategies.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If time threshold-based eviction is used, then cache eviction can be automated, but frequently used entries are incorrectly evicted

Engineering Contradiction:

Improvecache eviction automationVSAvoideviction accuracy

Core Design Contradiction:

Extent of automationVSReliability

Solution Approach 1:

The patent changes the parameter used for eviction decisions from fixed time thresholds to dynamic scores that incorporate multiple factors including access frequency, recency, and size. This allows the system to automatically evict entries while accurately identifying the least valuable ones, resolving the contradiction between automation and reliability.

Inventive Principle:

Principle #35Parameter changes

2Quantity of substance

If fixed size cache is used, then cache memory is efficiently managed, but frequently used entries are displaced by less important entries

Engineering Contradiction:

Improvecache memory utilizationVSAvoidcache hit rate

Core Design Contradiction:

Quantity of substanceVSProductivity

Solution Approach 1:

The patent introduces a scoring mechanism that evaluates multiple parameters (access frequency, recency, size) to determine eviction candidates. This ensures that within the fixed cache size, the entries with the lowest composite scores are evicted, preserving frequently used entries and maintaining high cache hit rates while efficiently utilizing cache memory.

Inventive Principle:

Principle #35Parameter changes

3Speed

If simple eviction criteria are used, then eviction process is fast, but eviction decisions are suboptimal

Engineering Contradiction:

Improveeviction process speedVSAvoideviction decision quality

Core Design Contradiction:

SpeedVSReliability

Solution Approach 1:

The patent pre-calculates and maintains score values for each cache entry based on access patterns, recency, and size. When eviction is needed, the system simply selects the entry with the lowest pre-computed score, making the actual eviction decision fast while the scoring mechanism ensures optimal decisions. This resolves the contradiction between speed and reliability.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12386744B2Ghost list cache eviction

Publication Date: 2025.08.12 DELL PROD LP

AI summary

One example method includes a cache eviction operation. Entries in a cache are maintained in an entry list that includes a recent list, a recent ghost list, a frequent list, and a frequent ghost list. When an eviction operation is initiated or triggered, timestamps of last access for the entries in the entry list are adjusted by corresponding adjustment values. Candidate entries for eviction are identified based on the adjusted timestamps of last access. At least some of the candidates are evicted from the cache.