Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Graph-native Optimization Framework for Complex Graph Queries

Bingqing Lyu, Xiaoli Zhou, Longbin Lai, Yufan Yang, Yunkai Lou, Wenyuan Yu, Jingren Zhou

TL;DR

GOpt addresses the challenge of efficiently processing complex graph patterns by introducing a graph-native, language-agnostic intermediate representation (GIR) and a dual-strategy optimization engine: RuleBasedStrategy for heuristic, rule-driven improvements and PatternStrategy for cost-based, internal ordering of graph operators. Its architecture decouples parsing, optimization, and execution, enabling seamless backend integration with Neo4j and GraphScope through a physical-converter interface and an optional protobuf-based plan format. The framework formalizes semantic variations in pattern matching, supports extensive optimization rules (e.g., FilterIntoPattern, EVFusion, ComSubPattern), and demonstrates substantial performance gains across a spectrum of queries, including single-pattern, multi-pattern, path, cyclic, and complex CGPs on SNB-like workloads. The work provides practical guidance for industrial graph analytics pipelines, showing how to apply hierarchical optimizations and cost-aware plan selection to achieve scalable, graph-native query processing.

Abstract

This technical report extends the SIGMOD 2025 paper "A Modular Graph-Native Query Optimization Framework" by providing a comprehensive exposition of GOpt's advanced technical mechanisms, implementation strategies, and extended evaluations. While the original paper introduced GOpt's unified intermediate representation (GIR) and demonstrated its performance benefits, this report delves into the framework's implementation depth: (1) the full specification of GOpt's optimization rules; (2) a systematic treatment of semantic variations (e.g., homomorphism vs. edge-distinct matching) across query languages and their implications for optimization; (3) the design of GOpt's Physical integration interface, enabling seamless integration with transactional (Neo4j) and distributed (GraphScope) backends via engine-specific operator customization; and (4) a detailed analysis of plan transformations for LDBC benchmark queries.

A Graph-native Optimization Framework for Complex Graph Queries

TL;DR

GOpt addresses the challenge of efficiently processing complex graph patterns by introducing a graph-native, language-agnostic intermediate representation (GIR) and a dual-strategy optimization engine: RuleBasedStrategy for heuristic, rule-driven improvements and PatternStrategy for cost-based, internal ordering of graph operators. Its architecture decouples parsing, optimization, and execution, enabling seamless backend integration with Neo4j and GraphScope through a physical-converter interface and an optional protobuf-based plan format. The framework formalizes semantic variations in pattern matching, supports extensive optimization rules (e.g., FilterIntoPattern, EVFusion, ComSubPattern), and demonstrates substantial performance gains across a spectrum of queries, including single-pattern, multi-pattern, path, cyclic, and complex CGPs on SNB-like workloads. The work provides practical guidance for industrial graph analytics pipelines, showing how to apply hierarchical optimizations and cost-aware plan selection to achieve scalable, graph-native query processing.

Abstract

This technical report extends the SIGMOD 2025 paper "A Modular Graph-Native Query Optimization Framework" by providing a comprehensive exposition of GOpt's advanced technical mechanisms, implementation strategies, and extended evaluations. While the original paper introduced GOpt's unified intermediate representation (GIR) and demonstrated its performance benefits, this report delves into the framework's implementation depth: (1) the full specification of GOpt's optimization rules; (2) a systematic treatment of semantic variations (e.g., homomorphism vs. edge-distinct matching) across query languages and their implications for optimization; (3) the design of GOpt's Physical integration interface, enabling seamless integration with transactional (Neo4j) and distributed (GraphScope) backends via engine-specific operator customization; and (4) a detailed analysis of plan transformations for LDBC benchmark queries.

Paper Structure

This paper contains 36 sections, 1 equation, 17 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. The Definitions
  4. Pattern Matching Semantics
  5. GIR Abstraction
  6. Data Model
  7. Operators
  8. Graph Operators
  9. Relational Operators
  10. Applications of GIR
  11. Architecture
  12. Query Parser
  13. GIR Optimizer
  14. Details of Optimization Strategies
  15. RuleBasedStrategy
  16. ...and 21 more sections

Figures (17)

  • Figure 1: An example of pattern matching semantics.
  • Figure 2: An example of joining two patterns under Edge-Distinct semantics.
  • Figure 3: The GIR Representations of Cypher queries. In the following, the MATCH_PATTERN operator in an GIR will be illustrated as a graph view for simplicity.
  • Figure 4: GOpt System Overview.
  • Figure 5: DAG of Optimization Process.
  • ...and 12 more figures

Theorems & Definitions (2)

  • Remark 2.1
  • Remark 3.1