Building Broker Architecture for Compressed Timeseries Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current web services platforms face delays in data presentation due to the need for raw timeseries data to be retrieved and processed in response to requests, leading to inefficiencies in query times for building management systems (BMS) that rely on relational databases for storing and processing timeseries data.
Innovation Solution
A method involving quantization, encoding, and integer compression of timeseries data, where differences between adjacent values are calculated and compressed, allowing for faster storage and retrieval by separating data into portions that can be accessed quickly, with the option to perform these operations on edge devices or cloud platforms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If raw timeseries data is stored in a relational database without significant organization or processing, then data storage simplicity is improved, but data retrieval and processing time increases significantly
Solution Approach 1:
The patent applies preliminary action by pre-processing timeseries data during ingestion through quantization, encoding, and compression operations. The data is transformed into a compressed format with calculated differences between adjacent values before storage, so that when retrieval occurs, the data is already optimized for fast access and processing, eliminating the need for time-consuming operations at query time.
Solution Approach 2:
The patent segments the timeseries data into compressed portions that can be independently stored and accessed. By dividing the data into compressed segments with differential encoding, the system enables selective retrieval of data portions without requiring access to entire datasets, significantly reducing retrieval time while maintaining storage efficiency.
2Quantity of substance
If data is compressed and processed during ingestion, then data storage size is reduced, but processing complexity during data collection increases
Solution Approach 1:
The patent applies parameter changes by transforming the data representation during ingestion through quantization (changing numerical precision), encoding (changing data format), and compression (changing storage density). These parameter transformations reduce the storage size of timeseries data while the processing complexity is managed through automated algorithms that operate efficiently during the data collection phase.
Data Source
AI summary
A brokering architecture is disclosed. The brokering architecture focuses on storing changes to the building system. Measurements are quantized to an integer representation and encoded so that time series data is stored in a series of changes of value. The encoded timeseries may be compressed using compression schemes that advantageously make use of small changes between adjacent stored values of the timeseries. Compressed timeseries values allow more data to be stored in RAM or other short-term storage and provide a more responsive system to control physical systems like buildings. Timeseries compression may be provided through extensions accessed through a structured data format on the message bus. Extensions can cause the execution of certain instructions based on the fields in the structured data and may be used to provide functionality in addition to timeseries compression including BACnet communications, predictions, and simulations.