Graph-Oriented Time Database Storage via Delta-Copy+Log
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional graph databases lack efficient storage techniques for time-based data, leading to substantial storage overload and increased query latency, particularly in scenarios with dynamic behavior like IoT networks, where temporary connections and rapid data changes are common.
Innovation Solution
The Delta-Copy+Log approach stores graph updates in disjoint time windows by representing differences (deltas) between snapshots instead of complete snapshots, reducing storage space and query execution time through a forward and backward querying mechanism, with Bloom filters optimizing data retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If complete snapshots of the graph are stored at each time window, then query execution time is reduced, but storage space requirements increase substantially
Solution Approach 1:
The patent segments the storage structure into two distinct parts: reference snapshots stored at intervals and incremental deltas stored for each time window. This segmentation allows the system to balance between storing complete graph states and storing only changes, thereby reducing overall storage requirements while maintaining efficient query capabilities.
Solution Approach 2:
The patent uses selective copying by storing complete reference snapshots only at certain intervals rather than at every time window. Between reference snapshots, only incremental deltas are stored. This copying strategy reduces storage space while still enabling reconstruction of complete graph states when needed for queries.
2Quantity of substance
If only incremental deltas are stored between snapshots, then storage space is reduced, but query execution time increases due to needing to load all previous updates
Solution Approach 1:
The patent performs preliminary action by pre-storing reference snapshots at strategic intervals before they are needed for queries. These reference snapshots serve as starting points for incremental reconstruction, eliminating the need to load all previous updates from the beginning when a query is executed.
Solution Approach 2:
The patent introduces reference snapshots as intermediary structures that mediate between complete graph states and incremental deltas. These intermediaries serve as anchor points that break the continuous chain of deltas into manageable segments, enabling efficient query execution by providing starting points for incremental reconstruction.
3Measurement precision
If reference snapshots are stored frequently, then query accuracy for time-based queries is improved, but storage overhead increases
Solution Approach 1:
The patent applies parameter changes by adjusting the frequency and distribution of reference snapshots based on query patterns and data characteristics. Rather than using a fixed interval, the system can dynamically modify the timing and spacing of reference snapshots to optimize between query accuracy and storage overhead.
Data Source
AI summary
A system for storing a graph-oriented time database, including a set of containers associated with adjacent time windows, a container containing a sequence of successive modifications of the graph over a time window with which it is associated. At least some of the containers, called reference containers, also contain a snapshot of the graph within the tine widow associated with the reference container, called reference snapshot. At least some other containers, called delta containers, contain a difference image between two snapshots of the graph, respectively associated with the start and with the end of the time window associated with the delta container.