Forecast-model-aware data storage for time series

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Forecasting models, especially complex domain-specific ones, face challenges in achieving near real-time decision support due to high computational demands and inefficient data access patterns, leading to performance issues in calculation and memory latency.

Innovation Solution

A system with multiple memory modules and a processor that identifies access patterns of forecast models to determine optimal storage layouts, enabling sequential access to time series data, thereby improving calculation performance and reducing memory latency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If complex forecast models with many parameters are used to achieve sufficient accuracy, then forecasting accuracy is improved, but calculation performance deteriorates

Engineering Contradiction:

Improveforecasting accuracyVSAvoidcalculation performance

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent changes the parameter of data storage organization from conventional formats to a format optimized for the forecast model's access patterns. By storing data in a sequential order that matches how the forecast model accesses it, the system improves calculation performance without changing the forecast model's parameters or accuracy.

Inventive Principle:

Principle #35Parameter changes

2Ease of manufacture

If conventional data storage formats are used, then data storage simplicity is maintained, but memory access efficiency deteriorates

Engineering Contradiction:

Improvedata storage simplicityVSAvoidmemory latency

Core Design Contradiction:

Ease of manufactureVSLoss of time

Solution Approach 1:

The system performs preliminary action by pre-organizing data in storage according to the forecast model's access patterns before the actual forecasting computation occurs. This preliminary organization of data in sequential order eliminates memory access delays during computation, as the data is already positioned for optimal access.

Inventive Principle:

Principle #10Preliminary action

3Adaptability or versatility

If data is stored in conventional formats, then storage compatibility is maintained, but sequential access capability deteriorates

Engineering Contradiction:

Improvestorage compatibilityVSAvoidsequential access speed

Core Design Contradiction:

Adaptability or versatilityVSSpeed

Solution Approach 1:

The patent applies dimensionality change by organizing data in a one-dimensional sequential storage format that aligns with the forecast model's access pattern, rather than using conventional two-dimensional or multi-dimensional storage structures. This dimensional reorganization enables continuous sequential access, improving access speed while maintaining storage compatibility through standardized data formats.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS10445399B2Forecast-model-aware data storage for time series data

Publication Date: 2019.10.15 SAP SE

AI summary

A system includes multiple memory modules arranged and configured to store data and at least one processor that is operably coupled to the memory modules. The at least one processor is arranged and configured to select an access pattern of a forecast model, determine a storage layout model based on the identified access pattern of the forecast model, and store values in an order defined by the storage layout model using at least one of the memory modules. The order of the stored values enables sequential access to the stored values for use in the forecast model. Implementations of one or more features of the system may be performed by a computer-implemented method and/or a computer program product.