1 BRVST: Efficient and Content-Expressive Information Matching Overlay in Wireless Networks Ying Li and Xin Wang Department of Electrical and Computer Engineering Stony Brook University Email: {yingli, xwang}@ece.sunysb.edu Abstract—Efficient and flexible information matching over wireless networks has become increasingly important and chal- lenging with the popularity of smart devices and the growth of social-network-based applications. Some existing approaches de- signed for wired networks are not applicable to wireless networks, due to their overwhelming control overheads. In this paper, we propose a reliable and scalable binary range vector summary tree (BRVST) infrastructure for flexible information expression support, effective content matching and timely information dis- semination over the dynamic wireless network. A novel attribute range vector structure has been introduced for efficient and accurate content representation and a summary tree structure to facilitate information aggregation. For robust and scalable operations over dynamic wireless network, the proposed overlay system exploits a virtual hierarchical geographic management framework. Extensive simulations demonstrate that BRVST has a significantly faster event matching speed, while incurs very low storage and traffic overhead, as compared with peer schemes tested. I. I NTRODUCTION With the drastic growth of social and wireless application information data generated and consumed, it is emergent to establish a bridge infrastructure that can timely and accurately discovers and delivers the information to various parties of interests. As an example of new era information service, a smart- phone user in a downtown block wants to obtain a recom- mendation for some restaurants while people close-by may be also searching for the same type of information. Another user just stepping out of a Thai cuisine is satisfied with the dining experience and would like to share this place with others. Other applications include traffic information posting and retrieval where users cooperatively contribute to and benefit from the real-time traffic reports. These applications can be better met by a "contribute-and- benefit" pattern system. Publish/Subscribe (Pub/Sub) system is one of this type, in which subscribers specify their inter- ests and publishers post advertisements. The system matches subscriptions with publications. Unlike client/server models, the Pub/Sub model decouples time, space, and flow between c 2014 IEEE. This paper has been accepted to MASS 2014. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, including reprinting/republishing this material for advertising or promotional purposes, collecting new collected works for resale or redistribu- tion to servers or lists, or reuse of any copyrighted component of this work in other works. publishers and subscribers to provide flexibility in information distribution. Gryphon [1] and SIENA [2] were once popular Pub/Sub models in wire-line networks, however, their tree-based struc- ture are not scalable in dynamic wireless network whose topology may constant change due to mobility and connection broken. Many later attempts have been made to apply Pub/Sub infrastructure for wireless networks [3] [4] [5], where the information in the systems is roughly divided into several basic types. These platforms cannot efficiently support het- erogeneous user application needs. Different from conventional Pub/Sub systems which mainly categorize information into a few types for ease of imple- mentation, the modern information system is expected to better meet the customized information needs of individual users. Besides the difference in categories, the heterogeneity of information is more generally resulted from different values or contents for the same type of information. In the restaurant recommendation example, the difference in the service time of a day or the average price level would totally distinguish restaurants and draw the interests of different groups of consumers, even when they provide the same type of foods. Simply ascribing information into coarse types (food, movie, car, etc.) cannot meet most application needs. On the other hand, completely expressing every detail of the information in words and matching over them is not feasible in reality. We need an information system that supports rich and accurate information content expression while efficiently reducing the representation complexity. In this paper, we propose a reliable and scalable content- expressive information matching and dissemination infrastruc- ture in a large-scale mobile wireless network, which utilizes novel and efficient components as well as a location-based virtual management infrastructure for efficient storage, light- weight communications, and quick information match. The main contributions of our work are: • We propose a mechanism to flexibly and efficiently represent information with the combination of a set of elementary tuples for numerical expression of the content. • We propose a novel Attribute Range Vector that allows flexible vector length adjustment based on the information accuracy requirement, and supports a unique simple bit- wise operation for quick content matching check, to facilitate accurate content representation as well as low- arXiv:1410.0706v1 [cs.NI] 2 Oct 2014