Time Series Data Clustering for Database Performance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for storing time series data in databases are inefficient, especially in high volume and high throughput scenarios, as they do not effectively manage large amounts of data and reduce performance during access operations.
Innovation Solution
The method involves clustering key figure values by time periods and partitioning data across multiple areas, allowing for efficient storage and retrieval through configurable clusters and data separation, which reduces memory requirements and improves performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If time series data is stored using a straightforward method with parameter values and time period as key, then the data structure is simple, but the performance deteriorates when accessing time series data among large amounts of data
Solution Approach 1:
The patent segments time series data into multiple data areas based on time periods, where each data area contains a subset of the time series data. This segmentation allows the system to access only relevant portions of data rather than scanning through all data, significantly improving access performance while maintaining manageable data structure complexity
Solution Approach 2:
The patent introduces a new dimension by organizing data into multiple data areas with hierarchical structure. Instead of a flat storage structure, it creates a multi-level organization where data areas are arranged in a hierarchy that enables efficient navigation and access, transforming the data access problem from linear search to hierarchical lookup
2Device complexity
If all time series data is stored in a single large table, then the data storage is simple, but the performance deteriorates due to large data volume during read and write operations
Solution Approach 1:
The patent divides the large time series data table into multiple smaller data areas, each handling a specific subset of data. This segmentation reduces the size of individual tables, improving read and write performance by operating on smaller data units while distributing the storage load across multiple areas
Solution Approach 2:
The patent implements dynamic data area assignment where data areas can be configured and adjusted based on access patterns and data characteristics. This dynamic organization allows the system to optimize storage and access efficiency by adapting the data area structure to specific operational requirements
3Device complexity
If data is not clustered by time periods, then the storage structure is simple, but the performance deteriorates as more database entries need to be accessed
Solution Approach 1:
The patent merges multiple time series data entries that share common characteristics (such as same data area and time period ranges) into a single database entry. This merging reduces the total number of database entries that need to be accessed, significantly decreasing data access time while the systematic organization maintains reasonable structure complexity
Data Source
AI summary
Methods and apparatus, including computer program products, are provided for storing data in a database. The data, whose structures are configurable, may comprise a time series, the time series comprising a set of key figure values of individual time periods, the key figure values being associated with a combination of parameter values, wherein the set of key figure values is organized in a number of configurable clusters.