DAGGER: A Scalable Index for Reachability Queries in Large Dynamic Graphs

TL;DR

DAGGER is a scalable, lightweight dynamic index for efficient reachability queries in large evolving graphs, suitable for real-world applications with millions of nodes.

Contribution

We introduce DAGGER, a novel fully dynamic reachability index based on interval labeling that scales to large, evolving graphs, outperforming baseline methods.

Findings

Effective on graphs with millions of nodes

Outperforms baseline methods in experiments

Suitable for real-world dynamic graph applications

Abstract

With the ubiquity of large-scale graph data in a variety of application domains, querying them effectively is a challenge. In particular, reachability queries are becoming increasingly important, especially for containment, subsumption, and connectivity checks. Whereas many methods have been proposed for static graph reachability, many real-world graphs are constantly evolving, which calls for dynamic indexing. In this paper, we present a fully dynamic reachability index over dynamic graphs. Our method, called DAGGER, is a light-weight index based on interval labeling, that scales to million node graphs and beyond. Our extensive experimental evaluation on real-world and synthetic graphs confirms its effectiveness over baseline methods.