Foundations and Trends® in Databases > Vol 7 > Issue 1-2

Big Graph Analytics Platforms

By Da Yan, The University of Alabama at Birmingham, USA, yanda@uab.edu | Yingyi Bu, Couchbase, Inc., USA, yingyi@couchbase.com | Yuanyuan Tian, IBM Almaden Research Center, San Jose, USA, ytian@us.ibm.com | Amol Deshpande, University of Maryland, USA, amol@cs.umd.edu

 
Suggested Citation
Da Yan, Yingyi Bu, Yuanyuan Tian and Amol Deshpande (2017), "Big Graph Analytics Platforms", Foundations and TrendsĀ® in Databases: Vol. 7: No. 1-2, pp 1-195. http://dx.doi.org/10.1561/1900000056

Publication Date: 12 Jan 2017
© 2017 D. Yan, Y. Bu, Y. Tian, and A. Deshpande
 
Subjects
Parallel and Distributed Database System,  Storage, Access Methods, and Indexing,  Distributed computing,  Parallel algorithms,  Programming Languages for Parallelism
 

Free Preview:

Download extract

Share

Download article
In this article:
1. Introduction
2. Preliminaries
I. Vertex-Centric Programming Model 
3. Vertex-Centric Message Passing (Pregel-like) Systems
4. Vertex-Centric Message-Passing Systems Beyond Pregel
5. Vertex-Centric Systems with Shared Memory Abstraction
II. Beyond Vertex-Centric Programming Model 
6. Matrix Algebra-Based Systems
7. Subgraph-Centric Programming Models
8. DBMS-Inspired Systems
III. Miscellaneous Issues 
9. More on Single-Machine Systems
10. Hardware-Accelerated Systems
11. Temporal and Streaming Graph Analytics
12. Conclusions and Future Directions
References

Abstract

Due to the growing need to process large graph and network datasets created by modern applications, recent years have witnessed a surging interest in developing big graph platforms. Tens of such big graph systems have already been developed, but there lacks a systematic categorization and comparison of these systems. This article provides a timely and comprehensive survey of existing big graph systems, and summarizes their key ideas and technical contributions from various aspects. In addition to the popular vertex-centric systems which espouse a think-like-a-vertex paradigm for developing parallel graph applications, this survey also covers other programming and computation models, contrasts those against each other, and provides a vision for the future research on big graph analytics platforms. This survey aims to help readers get a systematic picture of the landscape of recent big graph systems, focusing not just on the systems themselves, but also on the key innovations and design philosophies underlying them.

DOI:10.1561/1900000056
ISBN: 978-1-68083-242-6
116 pp. $99.00
Buy book (pb)
 
ISBN: 978-1-68083-243-3
116 pp. $270.00
Buy E-book (.pdf)
Table of contents:
1. Introduction
2. Preliminaries
I. Vertex-Centric Programming Model
3. Vertex-Centric Message Passing (Pregel-like) Systems
4. Vertex-Centric Message-Passing Systems Beyond Pregel
5. Vertex-Centric Systems with Shared Memory Abstraction
II. Beyond Vertex-Centric Programming Model
6. Matrix Algebra-Based Systems
7. Subgraph-Centric Programming Models
8. DBMS-Inspired Systems
III. Miscellaneous Issues
9. More on Single-Machine Systems
10. Hardware-Accelerated Systems
11. Temporal and Streaming Graph Analytics
12. Conclusions and Future Directions
References

Big Graph Analytics Platforms

The growing need to deal with massive graphs in real-life applications has led to a surge in the development of big graph analytics platforms. Tens of such big graph systems have already been developed, and more are expected to emerge in the near future. Although several experimental studies have been conducted in recent years that compare the performance of several big graph systems, Big Graph Analytics Platforms is the first text to provide a comprehensive survey that clearly summarizes the key features and techniques developed in existing systems. It aims to help readers get a systematic picture of the landscape of recent big graph systems, focusing not just on the systems themselves, but also on the key innovations and design philosophies underlying them. In addition to the popular vertex-centric systems which espouse a think-like-a-vertex paradigm for developing parallel graph applications, Big Graph Analytics Platforms also covers other programming and computation models, contrasts those against each other, and provides a vision for future research in the field.

Big Graph Analytics Platforms provides a timely and comprehensive survey of existing big graph systems, and summarizes their key ideas and technical contributions from various aspects. It is an ideal reference for researchers and practitioners already working, or about to embark on work in this growing area.

 
DBS-056