A Flexible Multi-Backend Framework for Temporal Graph Data Management
Abstract
Graph data structures are widely used to represent relationships among interconnected entities in modern applications such as social networks, recommendation systems, transportation networks, and communication infrastructures. Managing evolving graph data presents several challenges including efficient storage, temporal tracking, scalability, and query performance. This research proposes a modular multi-backend graph engine designed to support flexible storage integration and efficient graph processing. The proposed architecture separates graph processing logic from the storage layer through a storage abstraction interface that enables integration with multiple backend storage systems such as in-memory stores and persistent key–value databases. In addition, the system introduces a delta-based change tracking mechanism that records incremental updates rather than storing complete graph copies after every modification. A temporal snapshot manager maintains historical graph states, enabling time-based graph analysis and reconstruction of previous graph versions. Experimental observations demonstrate improvements in traversal efficiency, storage utilization, and scalability when managing dynamic graph datasets.