Catálogo de publicaciones - libros

A switched antenna system can provide transmission or reception in any desired direction by an array of directional antennas. Directional antennas have tremendous potential for improving the performance of wireless ad hoc networks[1]. While offering higher spatial reuse and larger transmission range, they also pose new challenges. Deafness is one of such problems, which arises when a transmitter fails to communicate to its intended receiver either because the receiver is beamforming towards a different direction[2]. As we have identified, generally, there might be three kinds of deafness problems. First, deafness-I happens when the intended receiver is a transmitter or receiver engaged in an ongoing transmission. Second, deafness-II occurs when the intended receiver lies in the area covered by an ongoing transmission and hence becomes deaf to the transmitter. Third, unlike the former two kinds of deafness which occur because RTSs cannot be heard by the intended receivers, deafness-III arises when the receiver has actually received RTS but cannot reply CTS, because it is aware of that this CTS will interfere with an ongoing transmission nearby. If left unaddressed, deafness problems not only severely degrade the performance at MAC layer but also considerably influence the upper-layer protocols, which would probably offset the benefits of directional antennas.

We introduce a new MANET structure model – party model – for the large scale MANET environment. For the party model MANET, we propose a new synthetic public key management scheme which applies web-of-trust and hierarchical trust simultaneously. The web-of-trust is used to design high efficient authentication between two nodes from the same party, and the frequent cooperation and communication inside a party can help to improve the security of authentication. In order to prevent falsification attack in remote authentication, we apply the hierarchical trust to establish CA in each party and use CA’s certificate as the trust intermediary. In the whole, our scheme can get a good tradeoff among security, overhead and flexibility.

In practice, it cannot easily be detected whether a computation running in a distributed system has terminated or not. Thus, suitable observing algorithms are required to solve this problem of .

We introduce a self-stabilizing algorithm that builds and maintains a spanning tree topology on any large scale system. We assume that the existing topology is a complete graph and that nodes may arrive or leave at any time. To cope with the large number of processes of a grid or a peer to peer system, we limit the memory usage of each process to a small constant number of variables, combining this with previous results concerning failure detectors and resource discovery.

The chase for the weakest system model that allows to solve consensus has long been an active branch of research in distributed algorithms. To circumvent the FLP impossibility in asynchronous systems, many models in between synchrony and asynchrony have been proposed over the years. Of specific interest is the chase for the weakest system model that allows the implementation of an eventual leader oracle Ω, and thus also enables consensus to be solved.

We present the first wait-free solution to dining philosophers under eventual weak exclusion in partially synchronous environments subject to crash faults. Potential applications include distributed daemon refinement for self-stabilizing algorithms.

We propose a self-stabilizing link formation algorithm based on a cooperative network formation game. An underlying network = (,) consists of processors represented by nodes = {1,2,...,}, and communication links represented by edges = { | , ∈ }. An agent network = (,), where = and ⊆, is defined on . Let be a benefit that agent provides others, and a cost of linking with that agent incurs. We assume a , a fair , and a for formation/severance of links.

”War of the worms” is a war between opposing computer worms, creating complex worm interactions. We propose a new Worm Interaction Model focusing on random-scan worm interactions. We validate our worm interaction model using extensive ns-2 simulations. This study provides the first work to characterize and investigate multiple worm interactions of random-scan worms in multi-hop networks. The main finding of this study is that maximum number of infected hosts can be drastically affected by the type of interaction.

We introduce a formal semantic-based modelling framework to model, specify and verify the functional and adaptive behaviour of synchronous adaptive systems.

Searching problem, which is to identify the peer that has some target resource, is an important and unique problem in Peer-to-Peer (P2P) file sharing systems. Since P2P file sharing systems maintain large and dynamic set of peers, the searching protocol is desired to be scalable and adaptive. To achieve such requirements, many kinds of search protocols are proposed. As one of those protocols, the searching protocol based on Probabilistic Weak Quorum System (PWQS) is recently proposed [1]. The principle of this protocol is as follows: In advance, a number of indices (location informations of an object) of each object are disseminated to randomly selected peers. When searching, the searcher sends a number of queries to randomly selected peers. If a query reaches a peer holding the index of the target object, search succeeds. It is shown that the protocol has advantages in the point of scalability, load balance and fault-tolerance.