Cloud Content Derivative Eviction Using Predictive Cost Modeling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Cloud-based content management systems face inefficiencies in managing storage costs due to the proliferation of both source and derivative content objects, with 'cold' objects being stored in high-performance appliances, leading to unnecessary expenses.

Innovation Solution

Applying a rule-based predictive model to identify candidate eviction objects and execute management commands to reduce storage costs while adhering to service level agreements, involving the use of a storage cost predictive model to rank and manage object evictions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Volume of stationary object

If additional high performance appliances are deployed to accommodate expanding content objects, then storage capacity increases, but storage costs increase

Engineering Contradiction:
Improvestorage capacityVSAvoidstorage costs
Core Design Contradiction:
Volume of stationary objectVSQuantity of substance

Solution Approach 1:

The system dynamically adjusts storage allocation by continuously monitoring access patterns and automatically transitioning content objects between hot and cold storage appliances based on their current activity status, ensuring storage capacity adapts to actual needs rather than static provisioning

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes the storage location parameter of content objects based on their access characteristics, moving frequently accessed objects to high performance appliances and infrequently accessed objects to lower cost appliances, thereby optimizing the cost-capacity tradeoff

Inventive Principle:
Principle #35Parameter changes

2Speed

If cold content objects are stored in high performance appliances, then access speed is maintained, but storage costs increase unnecessarily

Engineering Contradiction:
Improveaccess speedVSAvoidstorage costs
Core Design Contradiction:
SpeedVSQuantity of substance

Solution Approach 1:

The system applies different storage quality characteristics to different content objects based on their access patterns, placing only hot content objects in high performance storage while moving cold content objects to lower cost storage, ensuring each object receives appropriate storage quality rather than uniform high performance storage

Inventive Principle:
Principle #3Local quality

3Ease of operation

If derivative content objects are generated and stored for all source content objects, then content accessibility is improved, but storage capacity requirements increase

Engineering Contradiction:
Improvecontent accessibilityVSAvoidstorage capacity requirements
Core Design Contradiction:
Ease of operationVSVolume of stationary object

Solution Approach 1:

The system dynamically manages derivative content objects by generating them on-demand when source objects are accessed and removing them when no longer needed, rather than pre-generating and permanently storing all possible derivatives, thus reducing storage capacity requirements while maintaining accessibility

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3574418B1Management of cloud-based shared content using predictive cost modeling
Publication Date: 2025.11.26 BOX INC
  • EP3574418B1 patent drawingFigure 1A
  • EP3574418B1 patent drawingFigure 1B
  • EP3574418B1 patent drawingFigure 1C

AI summary

Systems and methods for managing content in a cloud-based service platform. A method embodiment operates over storage content objects stored in storage devices in a cloud-based shared content management system. The method commences upon identifying a source object and identifying derivative objects that are generated based properties of the source object. After a time, candidate eviction objects are identified. One form of analysis is performed over source objects and another form of analysis is performed over derivative objects. Derivative objects are classified using the analysis, which classification is used to determine object management commands associated with the derivative object such as to remove the derivative object from one storage location (e.g., in a high-performance storage filer) and relocate it to another (e.g., lower cost) storage location. Based on the analysis, a derivative object might be deleted completely and then regenerated at a later time if/when needed.