Catálogo de publicaciones - libros

Wireless ad hoc and sensor networks need to deal with unstable links. In practice the link quality between neighboring nodes fluctuates significantly over time. In this paper we evaluate the impact of topology control on routing performance. We propose a dynamic version of the XTC topology control algorithm. This simple and strictly local protocol removes unreliable and redundant links from the network. By means of physical experiments on an indoor mica2 testbed we study the beneficial effects of topology control on source routing, one of the most common routing schemes for ad hoc and sensor networks. In particular we compare the performance of source routing with and without topology control. Our results show that topology control reduces route failures, increases network throughput, and diminishes average packet delay.

Recent research shows that the traffic in public wireless networks is mostly transient and bursty. There is good reason to believe that ad-hoc traffic will follow the same pattern as its popularity grows. Unfortunately transient traffic generates route discoveries much more frequently than the long-term, constant-bit-rate traffic, causing network congestion problems for existing routing protocols. This paper describes the design of a new routing algorithm, called ECBR, that uses hybrid backbone routing in a manner that is well suited to workloads that include transient traffic. We explain three key features of our algorithm and demonstrate their roles in greatly improving the performance compared to existing reactive and backbone routing techniques.

In this article, we compare two self-organization and hierarchical routing protocols for networks. These two protocols apply the reverse approach from the classical one, since they use a reactive routing protocol inside the clusters and a proactive routing protocol between the clusters. We compare them regarding the cluster organization they provide and the routing that is then performed over it. This study gives an idea of the impact of the use of recursiveness and of the partition of the DHT on self-organization and hierarchical routing in ad hoc networks.

Location-based multicast protocol for mobile ad hoc networks is proposed in this paper. A network is divided into grids according to geographical location information and the grid network is divided into a high-channel subnetwork and a low-channel subnetwork according to labels of grids, where the destination set is divided into subsets according to its source node. Then destination nodes are partitioned into groups by using location information, the multicast routing is done in label order for each group. The proposed protocol does not require the maintenance of a distribution structure(e.g., a tree or a mesh). A forwarding node only uses information about positions of its destinations and its own neighbors to determine next hops that packet should be forwarded to and is thus very well suited for highly dynamic networks. Proposed protocol is scalable.

Location-based routing techniques, greedy routing and face routing, route data by using the location information of wireless nodes. Greedy routing efficiently routes data in dense networks by giving short hop paths, but it does not guarantee message delivery. Face routing has been designed and combined with greedy routing to achieve both transmission efficiency and guaranteed message delivery. The existing face routing algorithms mainly works on three types of planar graphs: Gabriel graph, relative neighborhood graph, and Delaunay triangulation. One major observation is that each transmission in face routing only can pass message over a short distance, resulting in that the existing face routing traverses long hop paths to destinations. In this paper, we present a Skip Face Routing (SFR) to reduce the face traversal cost incurred in the existing approaches. By using simulation studies, we show that SFR significantly increases routing performance.

Wireless sensor networks have attracted great attention in research and industrial development due to their fast-growing application potentials. Most of the existing geographic routing algorithms for wireless sensor networks are based on maintaining on the sensor node, leading to a well-known void problem. In this paper, we demonstrate that we can improve routing performance by storing more neighbors (called ) on the sensor nodes so as to avoid the void problem. We propose a neighborhood discovery algorithm and a neighborhood maintenance strategy to collect and maintain the for each node. Based on the on each node, we propose an algorithm. Simulation results show that the AVGR routing algorithm outperforms the typical routing algorithm GPG/GPSR, especially in networks with more voids.

The correctness of a routing protocol consists of two kinds of properties: safety and liveness. Safety properties specify that every route found by the protocol is well formed, while liveness properties specify that useful routes will eventually be found and data messages be eventually delivered to recipients. Many safety properties for routing protocols have been verified; however, the verification of liveness properties was overlooked. This paper stresses the importance of liveness properties of routing protocol and presents a formal verification of the DSR (Dynamic Source Routing) protocol dealing with both safety and liveness properties. The results are checked with Isabelle/HOL/Isar.

Ad-hoc and cellular networks have received great attention in recent years. To accommodate the large number of users and traffic over a large geographic area, cellular networks could take advantage of the infrastructure-less ad-hoc networks to provide extended service. One of the key issues in the integration of cellular and ad-hoc networks is to find some mobile nodes as proxies to relay the messages from base stations to destination nodes. In this paper, a scalable proxy relay routing protocol (SRP) is proposed to increase the total throughput of the system. In the strategy, the base station always sends data to the destination node through the selected proxy nodes which has minimum transmission delay. We demonstrate the advantage of our scheme over other current schemes such as UCAN and DST through simulation. The results have shown that the scheme outperforms other related schemes in terms of throughput and end-to-end delay.

This paper presents ORAL, a new on demand routing protocol in mobile ad hoc network, which incorporates the on demand source routing, local repair and MPLS like label switching to achieve higher throughput at the expense of reasonable overhead. ORAL separates the routing and forwarding explicitly: in the former process a label switch path and corresponding traversed node sequence, e.g., the source route, are constructed and maintained similar to DSR except the label signaling embedded in control packets. In the latter process, packet forwarding is based on label matching and switching rather than destination address or prefix searching and matching which is well-known computationally intensive. Integrating label and refined source route, ORAL can control the route table size equal to the amount of actively communicating nodes if no other optimization is deployed. Numerous simulation experiments show that ORAL achieves significant higher throughput in comparison with other protocols like DSR and AODV but with much less protocol overhead than AODV though a bit more than DSR.

A mobile ad-hoc network (MANET) is one consisting of a set of mobile hosts capable of communicating with each other without the assistance of base stations. It is necessary to use bandwidth effectively because MANET has limited bandwidth. In this paper, we propose SMGS in which each node estimates its expected life time and if its ELT is larger than that of current gateway it becomes a candidate node. When a source node establishes a path, in each grid the candidate node will take the route request and be a gateway node for the each source node. The node that is expected to stay the longest time in the grid is selected so that we can reduce frequent gateway handoff, packet loss, and handoff delay.