Distributed Data Mart Architecture for Heterogeneous Performance Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In complex computing environments, existing technologies face challenges in aggregating and analyzing large volumes of heterogeneous data from various sources to provide timely, customized performance and capacity statistics to support personnel, often resulting in delayed identification of potential issues that could lead to service disruptions.
Innovation Solution
A method that collects raw data from multiple systems, stores it in a unified repository, filters and formats it for easier management, merges it with historical data to create long-term reports, and forwards these reports to personal devices for real-time or near-real-time analysis and reporting, enabling quick identification of trends and resource adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If a large centralized computing resource and data repository are used to aggregate and analyze large volumes of heterogeneous data, then data aggregation capability is improved, but processing speed and timeliness deteriorate
Solution Approach 1:
The patent divides the centralized data processing system into multiple distributed data marts, each handling specific data subsets. This segmentation allows parallel processing across multiple nodes, maintaining comprehensive data aggregation capability while significantly improving processing speed through distributed computation architecture.
Solution Approach 2:
The patent introduces a temporal dimension to data processing by implementing real-time data streaming capabilities alongside historical data analysis. This allows the system to process and deliver insights across multiple time dimensions simultaneously, improving both aggregation depth and processing speed.
2Loss of information
If data is collected and processed centrally to ensure comprehensive analysis, then data completeness is improved, but delivery timeliness to support personnel deteriorates
Solution Approach 1:
The patent pre-processes and structures data in distributed data marts before analysis is needed, organizing heterogeneous data into standardized formats and pre-computing baseline metrics. This preliminary action ensures data completeness is maintained while enabling rapid query execution and timely delivery of insights when needed.
Solution Approach 2:
The patent introduces distributed data marts as intermediary structures between raw data sources and end-user analytics. These data marts act as buffers that pre-aggregate and structure data, ensuring completeness is preserved while enabling fast retrieval and analysis without requiring real-time centralized processing.
3Adaptability or versatility
If customized analyses are generated for specific support personnel, then analysis relevance is improved, but processing complexity increases
Solution Approach 1:
The patent implements local quality by allowing each distributed data mart to be customized for specific user roles, support personnel needs, and organizational units. Each data mart can apply local filtering, aggregation, and presentation rules tailored to its audience, providing highly relevant customized analyses without requiring complex centralized customization logic.
Solution Approach 2:
The patent applies partial customization by providing different levels of analysis detail and customization based on user needs. Rather than fully customizing every analysis for every user, the system provides appropriate levels of customization - from standardized reports for general users to highly customized views for specialized roles - reducing overall processing complexity while maintaining relevance.
Data Source
AI summary
A method and associated system for method for generating performance and capacity statistics that consists of a processor receiving statistical information from a set of monitoring entities that monitor characteristics of one or more computing resources. The processor formats the received statistics for storage in a raw-data repository, then filters and processes the data to extract data items necessary to generate predefined reports and to place the extracted data in a format consistent with historical information. The processor then merges the filtered, formatted data with the historical information to create an integrated history of the characteristics and stores the integrated history in a history repository. This integrated history is automatically translated into reports customized to requirements of particular support, personnel, which are then forwarded to local devices from which the support personnel may select, customize, and review the reports.