TY - GEN
T1 - Bridging the GAP
T2 - 1st ACM SIGMOD Joint International Workshop on Graph Data Management Experiences and Systems and Network Data Analytics, GRADES-NDA 2018
AU - Iyer, Anand Padmanabha
AU - Panda, Aurojit
AU - Venkataraman, Shivaram
AU - Chowdhury, Mosharaf
AU - Akella, Aditya
AU - Shenker, Scott
AU - Stoica, Ion
N1 - Funding Information:
We would like to thank the reviewers for their valuable feedback. In addition to NSF CISE Expeditions Award CCF-1730628, this research is supported in part by DHS Award HSHQDC-16-3-00083, and gifts from Alibaba, Amazon Web Services, Ant Financial, CapitalOne, Ericsson, Facebook, Google, Huawei, Intel, Microsoft, Scotiabank, Splunk and VMware.
Publisher Copyright:
© 2018 Copyright held by the owner/author(s).
PY - 2018/6/10
Y1 - 2018/6/10
N2 - While there has been a tremendous interest in processing data that has an underlying graph structure, existing distributed graph processing systems take several minutes or even hours to execute popular graph algorithms. However, in several cases, providing an approximate answer is good enough. Approximate analytics is seeing considerable attention in big data due to its ability to produce timely results by trading accuracy, but they do not support graph analytics. In this paper, we bridge this gap and take a first attempt at realizing approximate graph analytics. We discuss how traditional approximate analytics techniques do not carry over to the graph usecase. Leveraging the characteristics of graph properties and algorithms, we propose a graph sparsification technique, and a machine learning based approach to choose the apt amount of sparsification required to meet a given budget. Our preliminary evaluations show encouraging results.
AB - While there has been a tremendous interest in processing data that has an underlying graph structure, existing distributed graph processing systems take several minutes or even hours to execute popular graph algorithms. However, in several cases, providing an approximate answer is good enough. Approximate analytics is seeing considerable attention in big data due to its ability to produce timely results by trading accuracy, but they do not support graph analytics. In this paper, we bridge this gap and take a first attempt at realizing approximate graph analytics. We discuss how traditional approximate analytics techniques do not carry over to the graph usecase. Leveraging the characteristics of graph properties and algorithms, we propose a graph sparsification technique, and a machine learning based approach to choose the apt amount of sparsification required to meet a given budget. Our preliminary evaluations show encouraging results.
UR - http://www.scopus.com/inward/record.url?scp=85050271524&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85050271524&partnerID=8YFLogxK
U2 - 10.1145/3210259.3210269
DO - 10.1145/3210259.3210269
M3 - Conference contribution
AN - SCOPUS:85050271524
T3 - Proceedings of the 1st ACM SIGMOD Joint International Workshop on Graph Data Management Experiences and Systems (GRADES) and Network Data Analytics (NDA), GRADES-NDA 2018
BT - Proceedings of the 1st ACM SIGMOD Joint International Workshop on Graph Data Management Experiences and Systems (GRADES) and Network Data Analytics (NDA), GRADES-NDA 2018
A2 - Bhattacharya, Arnab
A2 - Fletcher, George
A2 - Roy, Shourya
A2 - Arora, Akhil
A2 - Larriba Pey, Josep Lluis
A2 - West, Robert
PB - Association for Computing Machinery, Inc
Y2 - 10 June 2018
ER -