Two Phase Data Retrieval Using Named Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current datacenter systems face inefficiencies in storing and retrieving metric data due to the need for extensive redesign when adding new components, and existing solutions either consume excessive computing resources or sacrifice storage and retrieval speed.
Innovation Solution
A method involving an index server that synchronously stores metric instances in a flattened format and an inventory server that asynchronously stores metric data in an unflattened format, using a resource description framework (RDF) model, allowing for efficient storage and retrieval with reduced computing resource usage and flexible onboarding of new components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If a NoSQL distributed database management system is used to store metric data, then storage capacity can be increased, but database redesign is required every time a new component is added, increasing human resource costs and system complexity
Solution Approach 1:
The system segments the database into two independent components: an index database that stores metric instance identifiers and metadata, and a wide-column store database that stores actual metric data. This segmentation allows each component to be independently scaled and modified without affecting the other, eliminating the need for complete database redesign when adding new components.
Solution Approach 2:
The patent introduces an intermediary layer consisting of metric instance identifiers and a mapping mechanism between the index database and the wide-column store database. This intermediary allows new components to be added by simply creating new identifier mappings without requiring structural changes to the underlying database schema.
2Adaptability or versatility
If alternative systems are used to avoid database redesign, then flexibility to add new components is improved, but computing resources such as cache resources are excessively consumed
Solution Approach 1:
The system applies local quality by storing only essential indexing information (metric instance identifiers and metadata) in the index database, while storing the actual metric data in the wide-column store database. This allows the index database to remain lightweight and efficient for query processing, while the wide-column store handles the bulk data storage, optimizing resource distribution.
Solution Approach 2:
The system performs preliminary action by pre-computing and storing metric instance identifiers and their mappings in the index database before actual metric data queries are executed. This preliminary structuring enables rapid query resolution without requiring extensive computing resources during runtime, as the mapping relationships are already established.
3Reliability
If metric data is stored in a traditional database format, then data integrity is maintained, but retrieval time increases and system performance decreases
Solution Approach 1:
The patent extracts the indexing function from the main database storage system by creating a separate index database that stores only metric instance identifiers and metadata. This extraction allows queries to be resolved quickly by looking up identifiers in the lightweight index database, while the actual metric data resides in the wide-column store, optimizing both retrieval speed and data integrity.
Data Source
AI summary
Method and systems for data retrieval is provided. A query is received to search for metric data corresponding to a component of a datacenter, the component of the datacenter identified by a metric instance. An index is searched for the metric instance, the index comprising the metric instance synchronously stored in a flattened format. Further, a slot identification corresponding to the metric instance is determined, the slot identification identifying a location of the metric data in an inventory. Based on the determined slot identification, metric data is retrieved from the inventory, the inventory comprising the metric data asynchronously stored in an unflattened format. Additionally, a query result comprising the metric data corresponding to the search is communicated.