TY - GEN
T1 - An efficient location-aware top-k subscription matching for publish/subscribe with boolean expressions
AU - Jiang, Hanhan
AU - Zhao, Pengpeng
AU - Sheng, Victor S.
AU - Xu, Jiajie
AU - Liu, An
AU - Wu, Jian
AU - Cui, Zhiming
N1 - Funding Information:
This work was partially supported by Chinese NSFC project (61472263, 61402312, 61402311), and the US National Science Foundation (IIS-1115417).
Publisher Copyright:
© Springer International Publishing Switzerland 2016.
PY - 2016
Y1 - 2016
N2 - Location-aware publish/subscribe (pub/sub) has attracted a lot of attentions with the booming of mobile Internet technologies and the rising popularity of smart-phones. Subscribers subscribe their interests with their locations as subscriptions, and publishers publish geo-information as events. Many state-of-art applications with a massive amount of geo-information, such as location-aware targeted advertising systems, face this situation. Existing related work mainly focuses on unstructured geo-textual information. However, many online-to-offline applications have enormous geo-information with different structured descriptions. To handle such structured information, a new type of location-aware pub/sub approach is needed. In this paper, we handle these subscriptions using boolean expressions. Since the number of publishers and subscribers can be enormous, it is extremely important to improve the matching effectiveness and efficiency of top-k query processing. In this paper, we develop a novel solution named RRt-trees. RRt-trees integrates Rt-tree and a predicate index structure together to return top-k best matched subscriptions from a great number of events. Our experimental results on synthetic and real-world datasets show that RRt-trees achieve better performance than baseline methods.
AB - Location-aware publish/subscribe (pub/sub) has attracted a lot of attentions with the booming of mobile Internet technologies and the rising popularity of smart-phones. Subscribers subscribe their interests with their locations as subscriptions, and publishers publish geo-information as events. Many state-of-art applications with a massive amount of geo-information, such as location-aware targeted advertising systems, face this situation. Existing related work mainly focuses on unstructured geo-textual information. However, many online-to-offline applications have enormous geo-information with different structured descriptions. To handle such structured information, a new type of location-aware pub/sub approach is needed. In this paper, we handle these subscriptions using boolean expressions. Since the number of publishers and subscribers can be enormous, it is extremely important to improve the matching effectiveness and efficiency of top-k query processing. In this paper, we develop a novel solution named RRt-trees. RRt-trees integrates Rt-tree and a predicate index structure together to return top-k best matched subscriptions from a great number of events. Our experimental results on synthetic and real-world datasets show that RRt-trees achieve better performance than baseline methods.
KW - Boolean expressions
KW - Location-aware pub/sub
KW - Top-k
UR - http://www.scopus.com/inward/record.url?scp=84962449663&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-32049-6_21
DO - 10.1007/978-3-319-32049-6_21
M3 - Conference contribution
AN - SCOPUS:84962449663
SN - 9783319320489
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 335
EP - 350
BT - Database Systems for Advanced Applications - 21st International Conference, DASFAA 2016, Proceedings
A2 - Navathe, Shamkant B.
A2 - Shekhar, Shashi
A2 - Wang, X. Sean
A2 - Wu, Weili
A2 - Du, Xiaoyong
A2 - Xiong, Hui
PB - Springer-Verlag
T2 - 21st International Conference on Database Systems for Advanced Applications, DASFAA 2016
Y2 - 16 April 2016 through 19 April 2016
ER -