Communication Dans Un Congrès Année : 2025

Growing Up HAL: Historic and Property Graph Queries

Résumé

Numerous applications, from social media to the banking industry, traffic management, and science, model application data as directed graphs. Fast graph analytics systems are implemented in-memory. Among existing systems, History Adjacency List (HAL) [2, 3] is distinguished by the ability to correctly answer queries even when updates, received in the graph database from multiple sources, arrive out of order, that is, impacted by variable transmission delays that may jeopardize query answer correctness in a system unaware of such issues. In this paper, we describe two recent extensions we brought to HAL, and which increase its functionalities. First, we demonstrate its capabilities to support historical queries, that is, queries that carry over the state of the dynamic graph at a given point in time. Second, we describe extensions we implemented to make HAL support arbitrary numbers of properties on nodes and edges, making it compatible with the needs of the popular LDBC benchmark [9], and present performance results of HAL on this benchmark.

Fichier principal
Vignette du fichier
LSGDA25_11.pdf (1001.83 Ko) Télécharger le fichier

Dates et versions

hal-05156390 , version 1 (10-07-2025)

Licence

Identifiants

  • HAL Id : hal-05156390 , version 1

Citer

Angelos-Christos Anadiotis, Muhammad Ghufran Khan, Ioana Manolescu. Growing Up HAL: Historic and Property Graph Queries. LSGDA 2025 - 4th International Workshop on Large Scale Graph Data Analytics (in conjunction with VLDB 2025), Sep 2025, London, United Kingdom. ⟨hal-05156390⟩
519 Consultations
221 Téléchargements

Partager

  • More