Pedro M. Ruiz

Approximating optimal multicast trees in wireless multihop networks

We study the problem of computing minimal cost multicast trees in multi-hop wireless mesh networks. This problem is known as the Steiner tree problem, and it has been widely studied in fixed networks. However, we show in this paper that in multi-hop wireless mesh networks, a Steiner tree is no longer offering the lowest bandwidth consumption. So, we re-formulate the problem in terms of minimizing the number of transmissions. We show that the new problem is also NP-complete and propose heuristics to approximate such trees. Our simulations results show that the proposed heuristics offer a lower cost than Steiner trees over a variety of scenarios.

Acknowledgment-Based Broadcast Protocol for Reliable and Efficient Data Dissemination in Vehicular Ad Hoc Networks

We propose a broadcast algorithm suitable for a wide range of vehicular scenarios, which only employs local information acquired via periodic beacon messages, containing acknowledgments of the circulated broadcast messages. Each vehicle decides whether it belongs to a connected dominating set (CDS). Vehicles in the CDS use a shorter waiting period before possible retransmission. At time-out expiration, a vehicle retransmits if it is aware of at least one neighbor in need of the message. To address intermittent connectivity and appearance of new neighbors, the evaluation timer can be restarted. Our algorithm resolves propagation at road intersections without any need to even recognize intersections. It is inherently adaptable to different mobility regimes, without the need to classify network or vehicle speeds. In a thorough simulation-based performance evaluation, our algorithm is shown to provide higher reliability and message efficiency than existing approaches for nonsafety applications.

Reliable and Efficient Broadcasting in Vehicular Ad Hoc Networks

Most of the envisioned services over vehicular networks need to deliver information to all vehicles inside a certain region. Several such broadcasting protocols have been reported so far, but surprisingly only one of them addresses the issue of intermittent connectivity. In this paper, we present a broadcast protocol which is suitable for a wide range of vehicular scenarios and traffic conditions. The protocol employs local position information acquired via periodic beacon messages. Beacons are used by cars to decide whether or not they belong to a connected dominating set (CDS). Vehicles in the CDS use shorter waiting period before possible retransmissions. Identifiers of circulated broadcast messages are added to beacons as piggybacked acknowledgements. When waiting timeout expires, vehicle retransmits if it has at least one neighbor which did not acknowledge circulated message with the last beacon, and sets a new waiting period. Our algorithm does not depend on any parameter or threshold which varies its operation. Despite its simplicity, the protocol is shown to provide high reliability and efficiency by means of a simulation-based performance evaluation. It also greatly outperforms the only competing algorithm we found in the literature which explicitly considers different mobility scenarios.

Bandwidth-Efficient Geographic Multicast Routing Protocol for Wireless Sensor Networks

We present geographic multicast routing (GMR), a new multicast routing protocol for wireless sensor networks. It is a fully localized algorithm that efficiently delivers multicast data messages to multiple destinations. It does not require any type of flooding throughout the network. Each node propagating a multicast data message needs to select a subset of its neighbors as relay nodes towards destinations. GMR optimizes the cost over progress ratio where the cost is equal to the number of neighbors selected for relaying and the progress is the overall reduction of the remaining distances to destinations. Such neighbor selection achieves a good tradeoff between the bandwidth of the multicast tree and the effectiveness of the data distribution. Our cost-aware neighbor selection is based on a greedy set merging scheme achieving a O(Dnmin(D,n)3) computation time, where n is the number of neighbors of current node and D is the number of destinations. As in traditional geographic routing algorithms, delivery to all destinations is guaranteed by applying face routing when necessary. Our simulation results show that GMR outperforms position based multicast in terms of cost of the trees and computation time over a variety of networking scenarios

GMR: Geographic Multicast Routing for Wireless Sensor Networks

We present geographic multicast routing (GMR), a new multicast routing protocol for wireless sensor networks. GMR manages to preserve the good properties of previous geographic unicast routing schemes while being able to efficiently deliver multicast data messages to multiple destinations. It is a fully-localized algorithm (only needs information provided by neighbors) and it does not require any type of flooding throughout the network. Each node propagating a multicast data message needs to select a subset of its neighbors as relay nodes towards destinations. GMR optimizes cost over progress ratio. The cost is equal to the number of selected neighbors, while progress is the overall reduction of the remaining distances to destinations. That is, the difference between distance from current node to destinations and distance from selected nodes to destinations. Such neighbor selection achieves a good trade-off between the cost of the multicast tree and the effectiveness of the data distribution. Our cost-aware neighbor selection is based on a greedy set merging scheme achieving a O(Dn min(D, n)3) computation time, where n is the number of neighbors of current node and D is the number of destinations. This is superior to the exponential computational complexity of an existing solution (PBM) which tests all possible subsets of neighbours, and to an alternative solution that we considered, tests all the set partitions of destinations. Delivery to all destinations is guaranteed by applying face routing when no neighbor provides advance toward certain destinations. Our simulation results show that GMR outperforms previous multicast routing schemes in terms of cost of the trees and computation time over a variety of networking scenarios. In addition, GMR does not depend on the use of any parameter, while the closest competing protocol has one parameter and remains inferior for all values of that parameter

Beacon-less geographic routing made practical: challenges, design guidelines, and protocols

Geographic routing has emerged as one of the most efficient and scalable routing solutions for wireless sensor networks. In traditional geographic routing protocols, each node exchanges periodic one-hop beacons to determine the position of its neighbors. Recent studies proved that these beacons can create severe problems in real deployments due to the highly dynamic and error-prone nature of wireless links. To avoid these problems, new variants of geographic routing protocols that do not require beacons are being proposed. In this article we review some of the latest proposals in the field of beacon-less geographic routing and introduce the main design challenges and alternatives. In addition, we perform an empirical study to assess the performance of beacon-based and beacon-less routing protocols using a real WSN deployment.

BOSS: Beacon-less On Demand Strategy for Geographic Routing inWireless Sensor Networks

Geographic routing (GR) algorithms, require nodes to periodically transmit HELLO messages to allow neighbors know their positions (beaconing mechanism). To reduce the control overhead due to theses messages, beacon-less routing algorithms have recently been proposed. However, existing beacon-less algorithms have not considered realistic physical layers. Therefore, those algorithms can not work properly in realistic scenarios. In this paper we present a new beacon-less routing protocol called BOSS. Its design is based on the conclusions of our open-field experiments using Tmote-sky sensors. BOSS is adapted to error-prone networks and uses a new delay function to reduce collisions and duplicate messages produced during the phase of selecting the next forwarder. Through extensive simulations we compare BOSS with BLR and CBF and show that our scheme is able to achieve almost perfect packet delivery ratio (like BLR) while having a low bandwidth consumption (even lower than CBF).

Drive and share: efficient provisioning of social networks in vehicular scenarios

Social Networks are one of the latest revolutions in networking, allowing users with common interests to stay connected and exchange information. They have enjoyed great success not only for traditional Internet users, but also for mobile users. Recent efforts are also being done to make social networks available within vehicles. However, to exploit social networks at their full potential in a vehicular context a number of technical challenges and design issues need to be faced. In this article, we analyze those challenges and present an innovative solution for providing Social services on the vehicle based on IP Multimedia Subsystem and Machine to Machine capabilities. To demonstrate the viability of the proposed scheme we present a social network service called Drive and Share which offers relevant information to vehicles using our proposed architecture.

Cluster-based OLSR extensions to reduce control overhead in mobile ad hoc networks

Proactive routing protocols for Mobile Ad Hoc Networks (MANETs) traditionally fail to scale up to large networks, since they generate a big amount of routing overhead. Based on OLSR, a proactive solution specifically designed for dense ad hoc networks, we develop a low overhead protocol called Clustered OLSR (C-OLSR). C-OLSR assumes that somehow the network is partitioned into clusters, and restricts the propagation of topology control messages inside every cluster. The generation and forwarding of inter-cluster topology information is based on the use of Multipoint Relays (MPRs) at the level of clusters. Through a simulation study, we show that C-OLSR outperforms OLSR both in terms of overhead generation and achievable throughput.

Internet connectivity for mobile ad hoc networks: solutions and challenges

The interconnection of mobile ad hoc networks to fixed IP networks is one of the topics receiving more attention within the MANET working group of the IETF as well as in many research projects funded by the European Union. Several solutions have recently been proposed, but at this time it is unclear which ones offer the best performance compared to the others. In addition to introducing the main challenges and design options that need to be considered, we perform a simulation-based evaluation aiming at providing some insight on the performance of these approaches. These simulation results have proven themselves valuable by showing that some of the most eye-catching features of the proposed approaches have practical performance issues which need to be enhanced.