Immutable Versioned Data Pipeline History Preservation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current data pipeline systems require significant manual effort for maintenance and lack dataset versioning, making it difficult to track and preserve data history, especially when updating datasets, which can lead to issues with incorrect data being overwritten and errors in data transformation.

Innovation Solution

A history preserving data pipeline system that maintains immutable and versioned datasets by storing build catalogs and transaction entries, allowing for the tracing of dataset data to its source and derivation program versions, and automating the build process through directed acyclic graphs of build dependencies.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If data transformation steps overwrite old dataset versions with new versions, then data processing efficiency is improved, but data history tracking capability deteriorates

Engineering Contradiction:
Improvedata processing efficiencyVSAvoiddata history tracking capability
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent creates a copy of the dataset versioning system by maintaining build catalog entries that reference transaction identifiers. Instead of physically storing multiple complete dataset versions, the system creates metadata copies (build catalog entries) that point to the actual data versions through transaction IDs, enabling history tracking without duplicating large data volumes.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The build catalog serves as an intermediary layer between the data storage system and the versioning requirements. It stores transaction identifiers and build dependency information, acting as a mediator that enables historical data tracking without requiring direct storage of multiple dataset versions, thus resolving the conflict between storage efficiency and versioning capability.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If manual configuration is used to maintain data pipeline systems, then system flexibility and adaptability are improved, but maintenance effort and time consumption increase

Engineering Contradiction:
Improvesystem flexibilityVSAvoidmaintenance effort
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system implements self-service by automatically maintaining build catalog entries and tracking dataset versions through transaction identifiers. The automated build process recursively creates build catalog entries that document data transformations, eliminating the need for manual configuration and reducing maintenance effort while preserving system adaptability.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The build catalog provides feedback about data lineage and transformation history by storing build dependency information. This automated feedback mechanism tracks which datasets depend on which source data and transformation steps, enabling the system to automatically manage complexity without manual intervention.

Inventive Principle:
Principle #23Feedback

3Reliability

If complete dataset versions are stored for history preservation, then data追溯 capability is improved, but storage requirements and system complexity increase

Engineering Contradiction:
Improvedata追溯 capabilityVSAvoidstorage system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

Instead of copying complete dataset versions, the patent copies only the essential metadata (transaction identifiers and build dependency information) in the build catalog. This lightweight copying approach enables full data追溯 capability by referencing original data versions without duplicating large data volumes, thus reducing storage requirements and system complexity.

Inventive Principle:
Principle #26Copying

Data Source

PatentEP3018553B1History preserving data pipeline system and method
Publication Date: 2021.11.03 PALANTIR TECHNOLOGIES INC
  • EP3018553B1 patent drawingFigure 1
  • EP3018553B1 patent drawingFigure 2
  • EP3018553B1 patent drawingFigure 3

AI summary

A history preserving data pipeline computer system and method. In one aspect, the history preserving data pipeline system provides immutable and versioned datasets. Because datasets are immutable and versioned, the system makes it possible to determine the data in a dataset at a point in time in the past, even if that data is no longer in the current version of the dataset.