Received August 9, 2016, accepted September 16, 2016, date of publication October 6, 2016, date of current version November 8, 2016. Digital Object Identifier 10.1109/ACCESS.2016.2615323 An Optimized Flow Allocation in Vehicular Cloud MEYSAM AZIZIAN 1 , SOUMAYA CHERKAOUI 1 , (Senior Member, IEEE), and ABDELHAKIM HAFID 2 1 INTERLAB Research Laboratory, Université de Sherbrooke, Sherbrooke, QC J1K 2R1, Canada 2 Network Research Labaratory, University of Montreal, Montreal, QC H3T 1J4, Canada Corresponding author: M. Azizian (meysam.azizian@usherbrooke.ca) This work was supported in part by the Canadian National Science and Engineering Research Council and in part by the Fonds de recherche du Québec–Nature et technologies. ABSTRACT In this paper, a vehicular cloud (VC) model is adopted where vehicles offer data as a service. We propose solutions for efficient data delivery based on transmission scheduling methods where vehicles gather data from their mounted sensors. This is done by first organizing vehicles into clusters, so that each cluster works as VC. A distributed D-hop cluster formation algorithm is presented to dynamically form vehicle clouds. The algorithm groups vehicles into non-overlapping clusters, which have adaptive sizes according to their mobility. VCs are created in such a way that each vehicle is at most D-hops away from a cloud coordinator (broker). Each vehicle chooses its broker based on relative mobility calculations within its D-hop neighbors. After cloud construction, a mathematical optimization scheduling algorithm is used to maximize throughput and minimize delay in delivering data from vehicles to their VC broker. Our proposed optimization model implements a contention-free-based medium access control where physical conditions of the channel are fully analyzed. Extensive simulations were performed for different scenarios to evaluate the performance of the proposed cloud formation and cloud-based transmission scheduling algorithms. Results show that VCs formed by our algorithms are more stable and provide higher data throughputs compared with others. INDEX TERMS Cloud formation, transmission scheduling, vehicular cloud, VANET, optimization. I. INTRODUCTION Nowadays, most vehicles have integrated computers and data processing units available as standard. New advances in vehicular technology have allowed vehicles to be more intelligent, provide a more pleasant driving experience, and avoid accidents. These new advances rely on the capability of vehicles to collect and process data available from their on-board sensors. Enclosed in new vehicular technology, is also vehicular communication. With embedded communi- cation, vehicles can interact with their environment to sup- port advanced safety applications. In fact, the US Federal Communications Commission (FCC) has allocated a 75 MHz spectrum in the 5.9 GHz band for Dedicated Short Range Communication (DSRC), specifically for vehicular com- munications [1], [2]. With their communication, sensing, and processing power, vehicle capabilities could, how- ever, exceed the sole needs of safety applications [3], [4]. More sophisticated applications, such as Vehicular Cloud Computing (VCC) could be operable in tandem [4]. Cloud computing makes an abstraction of the used access technology, and the used communication architecture, while maintaining the idea of service ubiquity [5]. Unlike cloud computing, Mobile Cloud Computing (MCC) was introduced to extend that ubiquity to mobile users [6]. VCC is a concept which constitutes the merging of MCC and Vehicular Ad-hoc Networks (VANET). Vehicles are a good platform for computing and communication which is potentially underutilized [7], [8]. VCC aims to make an effi- cient use of resources available in vehicles, such as comput- ing, sensing, and communication to provide useful services. However, achieving VCC does not come without challenges. In VC, vehicles are dynamic and consequently available resources too. New developments must be made to support such mobility and dynamicity in resources [4], [9], [10]. In this paper, we propose solutions to form VCs which provide efficient service delivery to outside users. We con- sider specifically data-as-a-service (DaaS), wherein data is collected from mounted sensors on vehicles. Data can be used to enable diverse applications such as real-time vehic- ular traffic engineering, weather analysis, police reports, emergency management, navigation, etc. To deliver services, mobile vehicles dynamically establish VCs, hence becoming 6766 2169-3536  2016 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. VOLUME 4, 2016