Multidimensional Space Mapping for Key-Value Store Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current key-value stores are inefficient in supporting searches on any combination of attributes, particularly non-primary attributes, due to their restrictive lookup operations, which limits their scalability and ability to handle structured and semi-structured data effectively.
Innovation Solution
A distributed key-value store system that maps multi-attribute objects into a multi-dimensional Euclidean space, allowing objects to be retrieved by searching one or more attributes, with a coordinator algorithm managing node distribution and hyperspace regions, and using hash-based or order-preserving functions for efficient object placement and retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If key-value stores use single-key lookup operations, then lookup efficiency is improved, but the ability to search by any attribute is lost
Solution Approach 1:
The patent applies dimensionality change by mapping objects from a single-key namespace into a multi-dimensional Euclidean space where each dimension corresponds to an attribute. This allows the system to maintain efficient lookup operations while enabling searches by any attribute combination, as objects can be located along any dimension axis rather than being constrained to a single key dimension.
Solution Approach 2:
The patent implements universality by creating a data structure that performs multiple functions: it supports efficient single-key lookups like traditional key-value stores, while simultaneously enabling multi-attribute searches across any combination of object attributes. The multi-dimensional mapping allows the same structure to serve both the efficiency requirements of key-value operations and the versatility requirements of attribute-based queries.
2Adaptability or versatility
If relational databases support general querying mechanisms, then search flexibility is improved, but system overhead increases and scalability decreases
Solution Approach 1:
The patent applies segmentation by dividing the data storage and query processing into independent dimensional axes. Each attribute becomes a separate dimension that can be queried independently, allowing the system to handle complex multi-attribute queries by processing each dimension separately rather than requiring complex joins and relationships as in relational databases. This segmentation reduces system overhead while maintaining query flexibility.
3Measurement precision
If key-value stores enumerate all objects for non-primary attribute searches, then search completeness is improved, but performance deteriorates
Solution Approach 1:
The patent resolves this contradiction by organizing objects in a multi-dimensional space where each attribute corresponds to a dimension. When searching by non-primary attributes, the system can directly navigate to the relevant dimension and locate matching objects without enumerating all objects. This dimensional organization ensures search completeness while maintaining efficient performance through direct navigation rather than full enumeration.
Data Source
AI summary
This invention relates to a system and methods for determining the placement of an object in a distributed key-value store by mapping the object to nodes in multidimensional hyperspace. A search function supports efficient object retrieval, even when the search query requests multiple objects and specifies them through non-primary keys. In response to a search query, the search is translated into hyperregions in the hyperspace to determine the set of nodes that hold the queried data object. The number of contacted nodes and the number of scanned objects are significantly reduced in comparison to prior art techniques.