Cloud Triplestore Schema Translation for Scalable Semantic Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current triple store technology lacks scalability in storage volume and compute performance, limiting the implementation of Knowledge Graphs for Machine Learning and Artificial Intelligence analytics.
Innovation Solution
Implementing a triple store over Cloud Analytical Data Store (CADS) using a Data Processing Application (DPA) that creates a schema native to CADS, translating semantic queries into CADS-native formats and leveraging its multi-node capabilities for efficient storage and computation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If current triple store technology is used to store Knowledge Graph data, then semantic query capabilities are provided, but scalability of storage volume and compute performance deteriorates
Solution Approach 1:
The patent introduces an intermediary layer that translates semantic queries (SPARQL) into SQL queries executable on cloud data warehouses. This mediator component enables compatibility between triple store data models and relational database infrastructure, allowing semantic query capabilities to be maintained while leveraging the scalable storage and compute resources of cloud data warehouses.
Solution Approach 2:
The patent creates a virtual copy or projection of triple store data structures over the cloud data warehouse infrastructure. By representing triple store data using relational tables with subject-predicate-object columns, the system replicates triple store functionality on top of scalable warehouse infrastructure, enabling both semantic querying and large-scale storage without requiring dedicated triple store hardware.
2Adaptability or versatility
If current triple store technology is used to store Knowledge Graph data, then semantic query capabilities are provided, but compute performance deteriorates
Solution Approach 1:
The patent replaces the mechanical query execution system of traditional triple stores with a optimized SQL-based execution engine running on cloud data warehouses. By translating SPARQL queries into SQL and leveraging the warehouse's distributed query processing infrastructure, the system achieves significantly higher compute performance while maintaining semantic query functionality.
Solution Approach 2:
The patent changes the operational parameters of query execution by running queries on a cloud data warehouse infrastructure rather than a traditional triple store engine. This parameter change includes utilizing the warehouse's distributed architecture, optimized query planners, and scalable compute resources, resulting in improved compute performance for semantic queries.
3Device complexity
If triple store data is stored on a single node, then data structure simplicity is maintained, but scalability and computation capacity deteriorates
Solution Approach 1:
The patent segments triple store data into relational tables that can be distributed across multiple nodes in a cloud data warehouse. By organizing data into subject-predicate-object tables that can be independently stored and processed, the system achieves both data structure simplicity and multi-node scalability, with each node handling a portion of the distributed data.
Data Source
AI summary
Systems and methods are disclosed for processing data. The system represents semantic data in the semantic data storage using a schema native to Cloud Analytical Data Store (CADS) based on data defining a semantic model. The system modifies the schema based on a detected change in the semantic model. The system writes semantic data into the CADS, wherein the semantic data is formatted according to the schema using at least one of: (a) bulk load, or (b) a sequence of write requests. The system receives a semantic query. The system translates the semantic query into a translated query in a CADS-native format, wherein the translated query is formatted according to the schema. The system causes the CADS to provide an answer to the translated query based on data contained in the CADS.