Multi-Level Hash Tree for Datastore Consistency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Maintaining consistency between redundant data items across multiple datastores is challenging due to the reliance on centralized master servers or separate computing applications, which consume significant resources and may disrupt active data storage operations.

Innovation Solution

The implementation of multi-level hash trees allows peer storage servers to communicate and identify differences between replicated data items by generating hash values based on attributes, enabling efficient comparison and updating of data items without a master server, thus maintaining data currency as a background process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If centralized master servers or separate computing applications are used to maintain data consistency, then data consistency is improved, but computing resource consumption increases significantly

Engineering Contradiction:

Improvedata consistencyVSAvoidcomputing resource consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

Storage servers autonomously generate multi-level hash trees from their stored data and participate in distributed consensus protocols to determine data currency. Each server independently computes hash values and contributes to the Merkle tree structure, eliminating the need for centralized master servers while maintaining data consistency through self-organized peer-to-peer verification.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent creates simplified copies of data verification through hash trees, where instead of comparing entire data sets between servers, each server generates a compact Merkle tree representation. These hash-based copies enable efficient consistency verification with minimal computational overhead compared to full data comparison, reducing computing resource consumption while ensuring data integrity.

Inventive Principle:

Principle #26Copying

2Reliability

If centralized master servers or separate computing applications are used to maintain data consistency, then data consistency is improved, but active data storage operations are disrupted

Engineering Contradiction:

Improvedata consistencyVSAvoiddata storage operation continuity

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The distributed hash tree verification operates continuously in the background without requiring storage servers to stop their primary data storage and retrieval operations. Servers maintain their normal read/write functions while simultaneously participating in consistency verification through asynchronous consensus protocols, ensuring uninterrupted data storage services.

Inventive Principle:

Principle #20Continuity of useful action

Solution Approach 2:

Multi-level hash trees are generated and maintained proactively as storage servers receive data updates, enabling consistency verification to occur before conflicts arise. The system continuously updates hash values and re-computes Merkle trees in advance, so when data currency determination is needed, the verification can proceed immediately without disrupting ongoing operations.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If traditional consistency checking methods are used, then data currency can be determined, but significant computing resources are required for analysis

Engineering Contradiction:

Improvedata currency determinationVSAvoidcomputing resource consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent divides the data verification process into hierarchical segments using multi-level hash trees. Instead of comparing all data items between servers, the Merkle tree structure segments verification into levels: root hash comparison first, then selective descent into child nodes only when discrepancies are found. This segmentation reduces the computational scope from O(n) full data comparison to O(log n) hash value comparisons in normal operations.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system transforms the verification parameter from comparing entire data sets to comparing compact hash values. By changing the verification metric from raw data content to cryptographic hash representations, the computational complexity decreases dramatically while maintaining verification accuracy, as hash values provide deterministic and collision-resistant data fingerprinting.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11663192B2Identifying and resolving differences between datastores

Publication Date: 2023.05.30 ORACLE INT CORP

AI summary

Techniques for efficiently maintaining consistency of data items across storage partitions are disclosed using a hierarchical multi-level hash tree. Copies of a data item may be associated with corresponding attributes that are used to generate hash values for the data item. Hash values of the attributes may then be used to label nodes in a multi-level hash tree. Differences between the replicated copies of a data item may be quickly identified by comparing hash values associated with successively lower peer nodes in corresponding hash trees. Once identified, systems may update versions of a data item that are no longer current.