Enabling efficient and robust network service for vehicular users

• Main Author: Zhang, Da
• Other Authors: Yeo Chai Kiat
• Format: Theses and Dissertations
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Computer science and engineering::Computer systems organization::Computer-communication networks; DRNTU::Engineering::Computer science and engineering::Information systems::Information storage and retrieval
• Online Access: http://hdl.handle.net/10356/52842
• Institution: Nanyang Technological University

As smart mobile devices become more and more popular, mobile users wish to be able to enjoy the network service from everywhere including the driving vehicles. Providing efficient and robust network service for vehicular users is challenging due to the unique characteristics of vehicles. In this thesis, the author focuses on two topics: providing efficient content-distribution service and robust routing service. For the content distribution service, the objective is to leverage the WiFi-based Access Points (APs) that can be exploited by vehicles on the roads to maximize the vehicular download performance. However, the high mobility of vehicles limits the duration of connection with APs, thus constraining the data transfer opportunities. The response latency and varying bandwidth between the AP and remote data-origin server can further waste the valuable connection duration. How to maximize the effective data transfer during vehicle-AP connection becomes the main problem to be solved. For the robust routing service, the goal is to protect the routing module on vehicles from being disrupted by malicious users. Although the routing module is one of the most important part to enable vehicle-to-vehicle communication, it is vulnerable to various types of attacks which are difficult to be detected or detected with high false positive rate under the high mobility environment. In the first work, the author proposes a content prefetching system, called Cooperative Content Distribution System for Vehicles (CCDSV), which operates upon a network of APs to collaboratively distribute contents to moving vehicles. The general idea can be described as follows: the requested data is prefetched onto the APs ahead along the driving trajectory of the requesting vehicle, which can then download the prefetched data with a high throughput when connection is established, without resorting to the remote server or being bottlenecked by the AP's backhaul link. CCDSV solves several important issues in a practical system, like the robustness to mobility prediction errors, limited storage resources of APs and prefetching overheads. CCDSV organizes the cooperative APs into a novel structure, namely, the contact map which is based on the vehicular contact patterns observed by APs. To fully utilize the wireless bandwidth provided by APs, a representative-based prefetching mechanism is proposed, in which a set of representative APs are carefully selected and then share their prefetched data with others. The selection process explicitly takes into account the AP's storage capacity, storage status, inter-APs bandwidth and traffic loads on the backhaul links. In addition to prefetching content before the requesting vehicle arrives, an AP also replicates into its storage the popular content items which have been requested in the past in order to maximize the vehicular download progress. In the second work, the author proposes such a replication system for vehicular users, which is quite different from the traditional system for static Web users. The transient connection period makes a vehicle hard to download the entire file requested and thus the content retrieval is usually across several APs. Such characteristic poses two problems: (1) replication in units of entire file may be inefficient in terms of resource utilization; (2) the actual contribution of an individual AP can be affected by the other correlated APs during the content retrieval. The proposed replication system explicitly solves these problems by taking into account the content popularity, vehicle-AP contact pattern and content availability among correlated APs. The replication system proposed here can work together with the prefetching system to augment the download performance for vehicular users. Both the above works perform optimization on the side of APs. In the third work, the author aims to consider the optimization on the side of vehicular users. In both the prefetching and replication systems above, the vehicular users utilize the opportunistically connected APs to serve their delay-tolerant requests by delaying the download from cellular networks for a while. During the delay, the vehicular users can benefit from the prefetching and replication systems running on the APs, and can also reduce the data burden on the cellular network by offloading traffic to WiFi-based APs. An important question here is: how long should the user wait for the offloading? Delaying a longer period brings more potential offloading opportunities (i.e. offloaded data volume), but makes the user more and more impatient. Furthermore, if no satisfactory amount of data can be offloaded through WiFi during the delay period, the user would think the delay is worthless and lose his enthusiasm in such offloading service. How to find the best time to hand-back the data transfer to the cellular networks from the offloading status, to achieve the optimal trade-off between the offloaded data volume and user satisfaction, is the focus of this work. The author proposes a model based on semi-Markov process to accurately predict the potential future offloaded volume. The user satisfaction is modeled as a satisfaction function of the delay period. Then a series of algorithms are proposed to find the optimal handing-back time from offloading. Besides the content distribution service, routing service is another key component for vehicular user to communicate with one another. In the final work, the author concentrates on increasing the robustness of routing module on vehicles against malicious behavior. The challenges include how to enhance detection accuracy when facing the highly dynamic characteristic of such networks, and how to distinguish malicious accusations under a totally autonomous structure. The author proposes Distributed Court System (DCS), a complete Intrusion Detection System that intends to solve these challenges in a low-cost and robust way. DCS does not have a centralized entity, but relies on the collaboration among the vehicles neighbouring the suspected one, to integrate information, improve the detection accuracy, and reject dissemination of malicious accusation. Through mathematical analysis and simulation, the proposed DCS is proven to be effective in a highly mobile and hostile network environment. The issues above are important for vehicular users to get acceptable level of network service. In the rest of this thesis, we will illustrate in detail the challenges and our proposed solutions.