Stack-Based Memory Processing for Analytics Performance

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current network analytics systems rely on slow heap data structures, leading to prolonged processing times for large datasets, making it difficult to generate real-time reports on content delivery metrics, which is crucial for content providers to monitor usage and quality of service effectively.

Innovation Solution

The method involves converting reference types to value types and storing them in non-mechanical memory, specifically using solid state memory and processing them using stack memory to summarize and store the information efficiently, thereby reducing processing time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If heap data structures are used for storing metric information, then flexibility in accommodating different data types is improved, but processing speed deteriorates

Engineering Contradiction:

Improveflexibility in accommodating data typesVSAvoidprocessing speed

Core Design Contradiction:

Adaptability or versatilityVSSpeed

Solution Approach 1:

The patent changes the fundamental parameter of data structure from heap-based to stack-based memory allocation. This parameter change transforms the memory management approach, enabling faster processing while maintaining the necessary flexibility through the stack's structured allocation mechanism.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the heap-based memory allocation mechanism with a stack-based mechanism. This substitution eliminates the need for complex heap management operations and enables significantly faster data processing while maintaining adaptability through the stack's inherent structure.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Quantity of substance

If databases are used to store metric information, then data storage capability is improved, but query processing time deteriorates

Engineering Contradiction:

Improvedata storage capabilityVSAvoidquery processing time

Core Design Contradiction:

Quantity of substanceVSLoss of time

Solution Approach 1:

The patent replaces traditional database systems with a stack-based memory structure for storing and querying metric information. This substitution eliminates database overhead and indexing mechanisms, enabling direct access to data structures stored in memory, thereby dramatically reducing query processing time while maintaining comprehensive data storage capability.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent fundamentally changes the storage and retrieval mechanism from database-based to memory-stack based. This parameter change in the underlying data structure enables O(1) access times for metric queries while maintaining the ability to store large volumes of metric data.

Inventive Principle:

Principle #35Parameter changes

3Quantity of substance

If spinning disk storage is used, then storage capacity is improved, but access speed deteriorates

Engineering Contradiction:

Improvestorage capacityVSAvoidaccess speed

Core Design Contradiction:

Quantity of substanceVSSpeed

Solution Approach 1:

The patent substitutes mechanical spinning disk storage with solid state memory and stack-based memory structures. This substitution eliminates mechanical seek operations and rotational delays, enabling dramatically faster data access speeds while maintaining adequate storage capacity through the combined use of fast memory technologies.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS9537733B2Analytics performance enhancements

Publication Date: 2017.01.03 BRIGHTCOVE INC

AI summary

A method and apparatus for processing metric information is disclosed in one embodiment. Metric information is gathered from a number of end users. At least some of the reference types are converted to value types and stored in non-mechanical memory. The value types are manipulated to summarize the metric information. The value types are processed using the stack instead of the heap.