Unified Database Model for Low Latency Time-Series Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Maintaining multiple different database systems for transactional and analytical data processing incurs high financial, storage, and integration costs, leading to increased time and compute resources for time-series queries, and often results in inaccurate sales forecasts due to outdated data analysis.

Innovation Solution

Implementing a single database system that stores temporal hierarchical data within data objects, eliminating the need for separate child records and allowing for efficient execution of both transactional and analytical queries, thereby reducing query runtime and improving accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple different database systems are used for transactional and analytical data processing, then data processing functionality is comprehensive, but system costs and integration overhead increase significantly

Engineering Contradiction:

Improvedata processing functionalityVSAvoidsystem integration overhead

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent combines OLTP and OLAP database functionalities into a single database system. The unified database simultaneously handles transactional operations and analytical queries, eliminating the need for separate database systems and reducing integration overhead while maintaining comprehensive data processing capabilities.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The single database system is designed to perform multiple functions: it can execute transactional operations (inserts, updates, deletes) and analytical queries (time-series analysis, sales forecasting) within the same system. This multi-functionality replaces the traditional separate OLTP and OLAP database architecture.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Reliability

If multiple different database systems are maintained, then specialized processing is achieved, but storage and processing costs increase

Engineering Contradiction:

Improvespecialized processingVSAvoidstorage and processing costs

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The patent merges the storage and processing resources of separate OLTP and OLAP databases into a single unified database system. This consolidation reduces the total quantity of storage and processing resources required while maintaining specialized processing capabilities through unified query execution.

Inventive Principle:

Principle #5Merging (Combining)

3Reliability

If separate child records are used to store temporal hierarchical data, then data integrity is maintained, but query runtime increases

Engineering Contradiction:

Improvedata integrityVSAvoidquery runtime

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent merges temporal hierarchical data from separate child records into the parent data object itself. By storing temporal data directly in the parent object, the system maintains data integrity while eliminating the need to query multiple child records, thereby reducing query runtime significantly.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent extracts temporal hierarchical data from separate child record structures and integrates it directly into the parent data object. This extraction eliminates the overhead of querying and joining child records while preserving the temporal data integrity through direct storage in the parent object.

Inventive Principle:

Principle #2Taking out (Extraction)

4Reliability

If periodic updates are used to synchronize databases, then data consistency is achieved, but data becomes outdated between updates

Engineering Contradiction:

Improvedata consistencyVSAvoiddata freshness

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent merges transactional data and analytical data into a single unified database, eliminating the need for periodic synchronization updates. This unified structure ensures that analytical queries always access the most current transactional data, maintaining both consistency and freshness without the delays inherent in periodic update mechanisms.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS11243966B1Data model design supporting low latency time-series volume planning and analysis

Publication Date: 2022.02.08 SALESFORCE INC

AI summary

Methods, systems, and devices supporting data storage are described. A database system may implement a hierarchical organization in which child data objects store hierarchical data for parent data objects. However, to support low latency time-series volume planning and analysis, the database system may additionally store a data object with a first set of data fields that includes object-specific data for the data object and a second set of data fields that includes hierarchical data (e.g., time-based hierarchical data organized into separate data fields for different time segments) for the data object. The database system may support both transactions and analytical queries using the data model. For example, the database system may receive a query indicating a time period for predictive analysis and may execute the query on the second set of data fields for the data object (e.g., without searching for child data objects storing the time-based hierarchical data).