Efficient and adaptive congestion control for heterogeneous delay-tolerant networks

Low-cost mobile personal clouds

“…This paper extends our previous work to focus on integrating suitable machine learning algorithms into the cloud, allowing the user to have rich real time interaction, describing how the cloud can sense and interact with the environment and how the data can be shared across multi-hop communication with the remote destinations. In order to improve reliability of data communication, we use P2P DTN for to support communication between multiple personal clouds as proposed in [32,33,34]. This enables multiple personal clouds to cooperate and communicate together without the need of infrastructure and thus preserving higher levels of privacy.…”

“…Emerging work on anticipatory mobile computing [13] argues that context prediction and anticipatory decisionmaking will be integral part of future mobile systems but identifies that processing costs can be high (suitable usually for the top of the range mobile devices) and that the challenge of real-world implementations of such systems which are almost non-existent (as most of the work is done via simulations). Delay tolerant and opportunistic networking approaches address the lack of continuous network connectivity [18,19,20,21,22,32,33,34] and were shown to have high efficiency and reliability as core data transfer protocols both mobile social and mobile ad hoc vehicular networks. Research in [32,33,34] has shown that both communication performance results and achieved levels of user privacy and anonymity of newly proposed adaptive opportunistic DTN protocols are high even in the presence of high numbers of malicious nodes.…”

“…Delay tolerant and opportunistic networking approaches address the lack of continuous network connectivity [18,19,20,21,22,32,33,34] and were shown to have high efficiency and reliability as core data transfer protocols both mobile social and mobile ad hoc vehicular networks. Research in [32,33,34] has shown that both communication performance results and achieved levels of user privacy and anonymity of newly proposed adaptive opportunistic DTN protocols are high even in the presence of high numbers of malicious nodes. [35,36,38] combine mobility of sensor sources, data carrier and clouds to enhance routing/forwarding protocol in terms of energy efficiency, security, connectivity and throughput.…”

Reputation Aware Obfuscation for Mobile Opportunistic Networks

“…Prioritisation of overlay nodes is based first on their reputation values, and then on connectivity patterns, and interest profile similarity, and anonymisation history. More detailed discussion on our obfuscation heuristics is given in the supplementary file but here we give a brief overview of three main classes of heuristics: 1) locally evaluated heuristics driven by the network topology and contact history analysis include nodes' "betweenness" centrality [1], [11], social "similarity" [1], [11], and tie strength relationships [1], [11], between the nodes; 2) heuristics driven by interest and user profile analysis include the degree of profiles' similarity nodes share when they meets based on the number of matched profile attributes versus the number of total attributes (L), denoted as LabelSimilarity.…”

“…San Francisco Cab Trace [2] has been shown to be the most challenging trace with very short connectivity durations [1], very long isolation time [1] and low number of neighbours [1] compared to other social traces (such as Sassy [7] and Infocom [14]). For example, SF Cab [2] traces exhibit predominantly short contact durations (a mean of 31sec) while Infocom 2006 displays substantially longer contact durations (a mean of 3min).…”