Hypertable Database System for Time-Series Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional database systems fail to simultaneously support high write rates and complex query patterns for time-series data, leading to a trade-off between scalability and query power, with traditional relational databases struggling with write performance and NoSQL databases lacking rich query language support and index capabilities.

Innovation Solution

A database system that uses a hypertable abstraction, partitioning data across servers and storage devices, allowing for standard SQL queries and supporting both scale-up and scale-out architectures, with chunk management that optimizes data distribution and query execution across multiple nodes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If traditional relational database systems are used to support complex queries, then query power is improved, but write performance deteriorates under high ingest rates

Engineering Contradiction:
Improvequery powerVSAvoidwrite performance
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent segments the database into multiple partitions distributed across different servers, with each partition handling a subset of the data. This segmentation allows the system to maintain relational database query capabilities while distributing write operations across multiple nodes, thereby improving both query power and write performance simultaneously.

Inventive Principle:
Principle #1Segmentation

2Productivity

If NoSQL databases are used to achieve horizontal scalability, then scalability is improved, but query language support deteriorates

Engineering Contradiction:
ImprovescalabilityVSAvoidquery language support
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent creates a database system that performs multiple functions: it supports standard SQL query language like traditional relational databases while simultaneously providing horizontal scalability like NoSQL databases. The hypertable abstraction layer enables the system to offer both relational query capabilities and distributed storage benefits in a single unified system.

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Ease of operation

If data is stored in a single location to simplify management, then ease of operation is improved, but scalability deteriorates

Engineering Contradiction:
Improvemanagement simplicityVSAvoidscalability
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent introduces a hypertable abstraction layer that acts as an intermediary between the user and the distributed storage system. Users interact with the database as if it were a single unified table, while the system automatically manages the distribution of data across multiple partitions and servers. This intermediary layer hides the complexity of distributed storage from users, maintaining ease of operation while enabling scalability.

Inventive Principle:
Principle #24Intermediary (Mediator)

4Productivity

If data is partitioned across multiple servers to improve scalability, then scalability is improved, but query complexity deteriorates

Engineering Contradiction:
ImprovescalabilityVSAvoidquery complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The database system automatically performs query routing and optimization without requiring user intervention. When a query is executed, the system automatically determines which partitions contain the relevant data and routes the query to those specific partitions, then aggregates the results. This self-service mechanism handles the complexity of distributed querying internally, allowing users to write simple queries while the system manages the distributed execution automatically.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP3563268B1Scalable database system for querying time-series data
Publication Date: 2022.09.14 TIMESCALE INC
  • EP3563268B1 patent drawingFigure 1
  • EP3563268B1 patent drawingFigure 2
  • EP3563268B1 patent drawingFigure 3

AI summary

A database system stores data as hypertables that represent partitioned database tables. Each hypertable comprises chunks of data that may be distributed across multiple locations, each location comprising at least a storage device. The database system provides an interface that allows database queries seamlessly to hypertables as well as standard tables. The database system dynamically creates chunks as records are added to a hypertable. The database system defines a new partitioning strategy if the storage configuration of the database system is changed by adding new locations or removing existing locations. The records added to the hypertable before the storage configuration was changed continue to be stored as chunks distributed according to the previous partitioning policy.