LeapGraph: A Fully External Graph Processing System on High-Speed SSD
As graphs exponentially grow nowadays, it becomes more and more challenging to hold large-scale graphs inside the memory of a single machine. To address this issue, fully external graph systems offer a promising solution of high scalability by keeping the large-scale graph in storage while loading a small part of the graph into memory for computation. However, the rapid evolution of storage technology presents a challenge for traditional fully external graph systems because they are unable to leverage these high-speed storage, thus severely limiting their performance. As a result, this work aims to present a fully external graph system, LeapGraph, developed for modern, high-speed storage with three main designs. Our preliminary evaluation with billion-scale graphs reveals the promising result of LeapGraph.
Authors
- Yuan-Hao Chang (ORCID: https://orcid.org/0000-0002-1282-2111)
- Tsun-Yu Yang (ORCID: https://orcid.org/0009-0002-8992-480X)
- Ming-Chang Yang (ORCID: https://orcid.org/0000-0002-4029-757X)
Institutions
- Chinese University of Hong Kong (HK)
- Institute of Information Science, Academia Sinica (TW)
Publication Details
- Published
- 2024-08-21
- DOI
- https://doi.org/10.1109/nvmsa63038.2024.10693656
- Primary Topic
- Advanced Data Storage Technologies
- Type
- article
- Field-Weighted Citation Impact
- 0.00