Search

Various issues on network routing in a mobile ad-hoc network (MANET) have been studied with outstanding results. However, there is still room for improvement with regard to data accessibility, especially when the number of nodes increases in the network. To enhance the data accessibility in a MANET, a few replica allocation schemes have been proposed. In these schemes, the nodes are partitioned into groups so that each group has as many different data items as possible. However, grouping in a MANET requires an enormous amount of communication cost, since the network topology is changed dynamically. In this paper, we propose a scalable replica allocation scheme in which the nodes are grouped based on the mobility of the nodes. Grouping in the proposed scheme is done with one-hop communications in a fully distributed manner. Within each group, replicas are allocated to the group nodes based on the estimated data access frequencies that are computed periodically by the allocator of the group. It turns out that grouping with mobility and utilizing the estimated access frequency information enhances the stability of the topology of each group as well as the data accessibility. We evaluate the proposed scheme by simulations with the Network Simulator NS-3 along with the analytical evaluations. The results of simulations and analytical evaluations demonstrate the effectiveness of the proposed scheme in terms of communication cost. The proposed scheme reduces the communication cost by 347.1 % on average over the existing replica allocation scheme, while achieving comparable data accessibility. [ABSTRACT FROM AUTHOR]

We study a forwarding strategy in opportunistic networks which are one of the most challenging networks among mobile ad-hoc networks. In opportunistic networks, a node does not have knowledge about the entire network topology, which is essential in the mobile ad-hoc network's forwarding strategy. Thus, node behavior is exploited to calculate future contact opportunities for forwarding a message. Utilizing social network analysis (e.g., similarity and centrality) has been proposed to improve the accuracy of the calculation task. This paper proposes a forwarding strategy based on an analysis of network topology. In the proposed strategy, each node takes a sequence of snapshots of its first-order neighbors during a warm-up period. Each node exchanges its own snapshots with each other, and then aggregates the snapshots in order to extract the network topology information. The extracted network topology is analyzed by social network analysis methods: compactness and algebraic connectivity. Forwarding decisions are made using the analysis of the features (compactness and algebraic connectivity). We present simulations using NS-2 and the home-cell community-based mobility model to show that the proposed forwarding strategy results in delay performances similar to the epidemic forwarding scheme, while maintaining reasonable network traffic. In addition, we demonstrate that the proposed strategy outperforms the SimBet and PRoPHET forwarding schemes with various communication ranges and memory space. [ABSTRACT FROM AUTHOR]

An opportunistic network (OPPNET) consists of diverse mobile nodes with various mobility patterns. Numerous mobility patterns and the resource constraints of mobile nodes lead to network partitioning that result in system performance degradation including low data accessibility. In a traditional mobile ad hoc network (MANET) which is similar to an OPPNET, replica allocation schemes have been proposed to increase data accessibility. Although the schemes are efficient in a MANET, they may not be directly applicable to an OPPNET because the schemes are based on a grouping of mobile nodes. It is very difficult to build groups based on network topology in an OPPNET because a node in an OPPNET does not keep its network topology information. In this paper, we propose a novel replica allocation scheme for an opportunistic network called the Snooping- based Fully Distributed replica allocation scheme. The proposed scheme allocates replicas in a fully distributed manner without grouping to reduce the communication cost, and fetches allocated replicas utilizing a novel candidate list concept to achieve high data accessibility. In the proposed scheme, a node can fetch replicas opportunistically based on the candidate list. Consequently, the proposed replica allocation scheme achieves high data accessibility while reducing the communication cost significantly. Extensive simulation results demonstrate that the proposed scheme reduces the communication cost and improves data accessibility over traditional schemes. [ABSTRACT FROM AUTHOR]