Virtual Co-Parent Keys for Document Information Tree Hot Spot Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Databases in certain usage scenarios, such as telecommunication infrastructures, face imbalanced data distribution leading to longer I/O request wait times due to 'hot spots' where a few parent entries have millions of associated child entries, causing inefficiencies in data retrieval and storage.
Innovation Solution
The generation of virtual co-parent keys, which redistribute child entries across partitions by computing a co-parent key that is prepended to the parent key, allowing for a more balanced distribution of data and reducing imbalances between partitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If data is organized in a hierarchical DIT structure with parent-child relationships, then data retrieval is simplified and structured, but data distribution becomes imbalanced causing hot spots with millions of child entries under few parent entries
Solution Approach 1:
The patent applies segmentation by introducing virtual co-parent keys that divide the child entries into multiple virtual partitions. Each child entry is associated with both its original parent key and a virtual co-parent key, effectively segmenting the imbalanced parent-child relationship into multiple balanced virtual relationships. This allows child entries to be distributed across multiple virtual partitions rather than concentrated under a single parent, reducing hot spots and I/O wait times while maintaining the hierarchical structure's operational simplicity.
2Device complexity
If child entries are concentrated under few parent entries to maintain hierarchical structure, then data organization is simplified, but partition imbalance occurs leading to inefficient data storage and retrieval
Solution Approach 1:
The patent introduces virtual co-parent keys as an intermediary mechanism between parent and child entries. These virtual co-parent keys act as mediators that redistribute child entries across multiple virtual partitions without altering the original parent-child hierarchical relationships. The virtual co-parent keys enable balanced data distribution across partitions while maintaining the simplified hierarchical organization, thus resolving the contradiction between organizational simplicity and retrieval efficiency.
3Loss of time
If virtual co-parent keys are generated to redistribute child entries across partitions, then data distribution becomes balanced reducing hot spots, but additional computational overhead is introduced for key generation and management
Solution Approach 1:
The patent implements self-service by making each child entry self-associate with its virtual co-parent key through a deterministic computation process. The virtual co-parent key is generated by computing a hash function on the child entry's distinguished name, allowing the system to automatically determine partition distribution without manual intervention. This self-service mechanism reduces I/O wait times through balanced distribution while minimizing key management complexity through automated, rule-based key generation.
Data Source
AI summary
Example implementations relate to generating a virtual co-parent key for a document information tree (DIT) including a plurality of child indexes each associated with a parent index. A virtual co-parent key for a first child index is generated. The first child index may be mapped to a child entry and may have a unique identification key including a child key and associated parent key. Generation of the co-parent key may include the computation of a co-parent seed key from the parent key or child key of the child index. Generation of the co-parent key may further include pre-pending the unique identification key with the computed co-parent seed key.