Compiler Infrastructure for Multidimensional Metadata Transformation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing OLAP systems face challenges in efficiently transforming multidimensional analytical metadata from various data sources into a format executable by in-memory computing engines, due to differing parameters and semantics across application servers, requiring a generic transformation platform to facilitate cross-database operations.
Innovation Solution
A compiler infrastructure is generated by transforming multidimensional analytical metadata into in-memory computing engine executable metadata, using a transformation library and pattern generator to map metadata with calculation patterns, enabling execution across various application servers without rewriting metadata, and deploying the calculation scenario in the in-memory computing engine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a generic transformation platform is implemented to transform multidimensional analytical metadata from various data sources, then adaptability across different application servers is improved, but device complexity increases
Solution Approach 1:
The patent introduces a transformation platform as an intermediary layer between multidimensional analytical metadata from various data sources and the in-memory computing engine. This intermediary transforms and standardizes metadata from different application servers into a unified format, enabling adaptability without requiring direct integration with each data source, thus managing complexity through abstraction.
Solution Approach 2:
The transformation platform is designed as a universal system that can handle metadata from multiple different application servers and data sources through a single unified interface. By implementing generic transformation rules and patterns, the system achieves multi-functionality, allowing one platform to serve multiple purposes across different data sources without requiring separate specialized systems for each.
2Ease of operation
If metadata transformation is performed to enable cross-database operations, then ease of operation across different data sources is improved, but loss of time in transformation process increases
Solution Approach 1:
The system performs preliminary transformation of multidimensional analytical metadata into a standardized format during the compilation phase, before the actual analytical processing begins. By preparing and standardizing the metadata structure in advance through transformation rules and patterns, the system reduces transformation time during execution while maintaining ease of operation across different data sources.
3Productivity
If in-memory computing engine is used for analytical processing, then productivity of analytics is improved, but use of energy increases
Solution Approach 1:
The patent extracts and transforms only the necessary multidimensional analytical metadata into the in-memory computing engine, rather than loading entire datasets. By selectively extracting and transforming only the required metadata structures and calculation patterns, the system achieves high analytics productivity while minimizing energy consumption compared to loading complete data into memory.
Data Source
AI summary
In an embodiment, the compiler infrastructure allows execution of multidimensional analytical metadata from various databases by providing a generic transformation. A compilation request to execute a multidimensional analytical metadata is received. A type of the compilation request is determined to identify an associated transformation and corresponding transformation rules. Based upon the type of compilation request, a database of an application server is queried to retrieve the corresponding multidimensional analytical metadata. Based upon the identified transformation rules, the multidimensional analytical metadata is transformed into a generic metadata that is executable by any desired engine. An instance of a calculation scenario is generated based upon the transformation. The compiler infrastructure is generated by deploying the instance of the calculation scenario in the desired engine (e.g. in-memory computing engine.)