TurboLynx at VLDB 2026

Saqib Ali · 2026-06-21T21:57:20+00:00

TurboLynx is an MIT-licensed, open-source property graph DBMS built for high-throughput analytical queries on schemaless graphs without ETL, using cost-based columnar clusters called Graphlets to efficiently store heterogeneous attributes while avoiding schema rigidity. Delivered as a lightweight, embeddable library with no server or daemon, it supports native Cypher execution and enables complex analytical workloads-combining multi-hop traversals, filtering, grouping, and top-k operations in a single pipeline. Benchmarks report that TurboLynx outperforms Neo4j in multi-hop traversal and DuckDB in graph-shaped joins, with full compliance across LDBC (464/464), TPC-H (22/22), and Neo4j IC1-IC14 verification tests, and its architecture has been accepted for presentation at VLDB 2026. Read more: https://www.linkedin.com/feed/update/urn:li:activity:7473622224323694592/

From: Saqib Ali | Date: Sun, 21 Jun 2026 21:57:20 +0000 #
Subject: TurboLynx at VLDB 2026
TurboLynx is an MIT-licensed, open-source property graph DBMS built for high-throughput analytical queries on schemaless graphs without ETL, using cost-based columnar clusters called Graphlets to efficiently store heterogeneous attributes while avoiding schema rigidity. Delivered as a lightweight, embeddable library with no server or daemon, it supports native Cypher execution and enables complex analytical workloads-combining multi-hop traversals, filtering, grouping, and top-k operations in a single pipeline. Benchmarks report that TurboLynx outperforms Neo4j in multi-hop traversal and DuckDB in graph-shaped joins, with full compliance across LDBC (464/464), TPC-H (22/22), and Neo4j IC1-IC14 verification tests, and its architecture has been accepted for presentation at VLDB 2026. Read more: https://www.linkedin.com/feed/update/urn:li:activity:7473622224323694592/
[ Reply ]

[ Back to Group: comp.databases.turbolynx ]

Keyboard Shortcuts: [ j/n next | k/p prev | g top | r reply | u up ]