Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Accelerating Regular Path Queries over Graph Database with Processing-in-Memory

Ruoyan Ma, Shengan Zheng, Guifeng Wang, Jin Pu, Yifan Hua, Wentao Wang, Linpeng Huang

TL;DR

Evaluation of Moctopus demonstrates superiority over the state-of-the-art traditional graph database, as well as PIM-friendly dynamic graph partitioning algorithm, which tackles graph skewness and preserves graph locality with low overhead for RPQ processing.

Abstract

Regular path queries (RPQs) in graph databases are bottlenecked by the memory wall. Emerging processing-in-memory (PIM) technologies offer a promising solution to dispatch and execute path matching tasks in parallel within PIM modules. We present Moctopus, a PIM-based data management system for graph databases that supports efficient batch RPQs and graph updates. Moctopus employs a PIM-friendly dynamic graph partitioning algorithm, which tackles graph skewness and preserves graph locality with low overhead for RPQ processing. Moctopus enables efficient graph update by amortizing the host CPU's update overhead to PIM modules. Evaluation of Moctopus demonstrates superiority over the state-of-the-art traditional graph database.

Accelerating Regular Path Queries over Graph Database with Processing-in-Memory

TL;DR

Evaluation of Moctopus demonstrates superiority over the state-of-the-art traditional graph database, as well as PIM-friendly dynamic graph partitioning algorithm, which tackles graph skewness and preserves graph locality with low overhead for RPQ processing.

Abstract

Regular path queries (RPQs) in graph databases are bottlenecked by the memory wall. Emerging processing-in-memory (PIM) technologies offer a promising solution to dispatch and execute path matching tasks in parallel within PIM modules. We present Moctopus, a PIM-based data management system for graph databases that supports efficient batch RPQs and graph updates. Moctopus employs a PIM-friendly dynamic graph partitioning algorithm, which tackles graph skewness and preserves graph locality with low overhead for RPQ processing. Moctopus enables efficient graph update by amortizing the host CPU's update overhead to PIM modules. Evaluation of Moctopus demonstrates superiority over the state-of-the-art traditional graph database.
Paper Structure (16 sections, 6 figures, 1 table)

This paper contains 16 sections, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. background
  3. Graph Database and Graph Partition
  4. PIM architecture
  5. Matrix-based graph operations
  6. design
  7. System architecture
  8. Graph partition
  9. locality-aware node distribution
  10. greedy-adaptive load balancing
  11. Optimizations for graph storage
  12. evaluation
  13. Evaluation Setup
  14. Performance of RPQs
  15. Performance of graph update
  16. ...and 1 more sections

Figures (6)

  • Figure 1: The architecture of the Moctopus with $P$ PIM modules.
  • Figure 2: Example of partitioning a routing connection graph in property graph and adjacency matrix view, and the matrix-based execution plan of a batch 2-hop path query.
  • Figure 3: Heterogeneous graph storage for high-degree nodes.
  • Figure 4: Run-time of $k$-hop path queries (log scale).
  • Figure 5: IPC cost of Moctopus and PIM-hash processing $3$-hop path queries (log scale).
  • ...and 1 more figures