Yingxia Shao - Large-scale Graph Analysis: System, Algorithm and Optimization

The big data paradigm presents a number of challenges for university curricula on big data or data science related topics. On the one hand, new research, tools and technologies are currently being developed to harness the increasingly large quantities of data being generated within our society. On the other, big data curricula at universities are still based on the computer science knowledge systems established in the 1960s and 70s. The gap between the theories and applications is becoming larger, as a result of which current education programs cannot meet the industrys demands for big data talents.

- Present a systematic and comprehensive knowledge structure for big data and data science research and education

More information about this series at http://www.springer.com/series/15869 - Supply lectures on big data and data science education with timely and practical reference materials to be used in courses

- Provide introductory and advanced instructional and reference material for students and professionals in computational science and big data

- Familiarize researchers with the latest discoveries and resources they need to advance the field

- Offer assistance to interdisciplinary researchers and practitioners seeking to learn more about big data

The scope of the series includes, but is not limited to, titles in the areas of database management, data mining, data analytics, search engines, data integration, NLP, knowledge graphs, information retrieval, social networks, etc. Other relevant topics will also be considered.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.

The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore

In this book, we will introduce readers to a methodology for scalable graph algorithm optimization in graph computing systems. Although the distributed graph computing system has been a standard platform for large graph analysis from 2010, it cannot efficiently handle advanced graph algorithms, which have complex computation patterns like dynamic and imbalance workload, huge amount of intermediate data, graph mutation, etc. Efficient and scalable large-scale graph analysis in general is a highly challenging research problem. We focus on the workload perspective and introduce a workload-aware cost model in the context of distributed graph computing systems. The cost model guides the development of high-performance graph algorithms. Furthermore, on the basis of the cost model, we subsequently present a system-level optimization resulting in a partition-aware graph-computing engine PAGE and present three efficient and scalable optimized graph algorithmsthe subgraph enumeration, graph extraction, and cohesive subgraph detection.

This book offers a valuable reference guide for junior researchers, covering the latest advances in large-scale graph analysis, and for senior researchers, sharing state-of-the-art solutions of advanced graph algorithms. In addition, all the readers will find a workload-aware methodology for designing efficient large-scale graph algorithms.