This paper studies the throughput-delay performance tradeoff in large-scale wireless ad hoc networks. It has been shown that the per source-destination pair throughput can be improved from Θ(1/√{nlogn}) to Θ(1) if nodes are allowed to move and a two-hop relay scheme is employed. The price paid for such a throughput improvement is large delay. Indeed, the delay scaling of the two-hop relay scheme is Θ(nlogn) under the random walk mobility model. In this paper, coding techniques are used to improve the throughput-delay tradeoff for mobile wireless networks. For the random walk mobility model, the delay is reduced from Θ(nlogn) to Θ(n) by employing a maximum distance separable Reed-Solomon coding scheme. This coding approach maintains the diversity gained by mobility while decreasing the delay.

We consider large-scale wireless sensor networks with n nodes, out of which k are in possession, (e.g., have sensed or collected in some other way) k information packets. In the scenarios in which network nodes are vulnerable because of, for example, limited energy or a hostile environment, it is desirable to disseminate the acquired information throughout the network so that each of the n nodes stores one (possibly coded) packet so that the original k source packets can be recovered, locally and in a computationally simple way from any k(1 + ¿) nodes for some small ¿ > 0. We develop decentralized Fountain codes based algorithms to solve this problem. Unlike all previously developed schemes, our algorithms are truly distributed, that is, nodes do not know n, k or connectivity in the network, except in their own neighborhoods, and they do not maintain any routing tables.

This paper studies the problem of resilience to node failures in large-scale networks modelled by random geometric graphs. Adopting a percolation-based viewpoint, the paper investigates the ability of the network to maintain global communication in the face of dependent node failures. Degree-dependent site percolation processes on random geometric graphs are examined, and the first known analytical conditions are obtained for the existence and non-existence, respectively, of a large connected component of operational network nodes after degree-dependent node failures. In electrical power networks or wireless communication and computing networks, cascading failure from power blackouts or virus epidemics may result from a small number of initial node failures triggering global failure events affecting the whole network. With the use of a simple but descriptive model, it is shown that the cascading failure problem is equivalent to a degree-dependent percolation process. The first analytical conditions are obtained for the occurrence and non-occurrence of cascading failures, respectively, in large-scale networks with geometric constraints.

We study the throughput-delay performance trade-off in large-scale wireless ad hoc networks. It has been shown that the per source-destination pair throughput can be improved from Θ(1/√n log n) to Θ(1) if nodes are allowed to move and a 2-hop relay scheme is employed. The price paid for such an improvement on throughput is large delay. Indeed, the delay scaling of the 2-hop relay scheme is Θ(n log n) under the random walk mobility model. In this paper, we employ coding techniques to improve the throughput-delay trade-off for mobile wireless networks. For the random walk mobility model, we improve the delay from Θ(n log n) to Θ(n) by employing Reed-Solomon (RS) codes. Our approach maintains the diversity gained by mobility while decreasing the delay.

We study the problem of wireless network resilience to node failures from a percolation-based perspective. In practical wireless networks, it is often the case that the failure probability of a node depends on its degree (number of neighbors). We model this phenomenon as a degree-dependent site percolation process on random geometric graphs. In particular, we obtain analytical conditions for the existence of phase transitions within this model. Furthermore, in networks carrying traffic load, the failure of one node can result in redistribution of the load onto other nearby nodes. If these nodes fail due to excessive load, then this process can result in a cascading failure. Using a simple but descriptive model, we show that the cascading failure problem for large-scale wireless networks is equivalent to a degree-dependent site percolation on random geometric graphs. We obtain analytical conditions for cascades in this model. This work represents the first investigation of cascading phenomena in networks with geometric constraints.

We study connectivity and transmission latency in wireless networks with unreliable links from a percolation-based perspective. We first examine static models, where each link of the network is functional (active) with some probability, independently of all other links, where the probability may depend on the distance between the two nodes. We obtain analytical upper and lower bounds on the critical density for phase transition in this model. We then examine dynamic models, where each link is active or inactive according to a Markov on- off process. We show that a phase transition also exists in such dynamic networks, and the critical density for this model is the same as the one for static networks under some mild conditions. Furthermore, due to the dynamic behavior of links, a delay is incurred for any transmission even when propagation delay is ignored. We study the behavior of this transmission delay and show that the delay scales linearly with the Euclidean distance between the sender and the receiver when the network is in the subcritical phase, and the delay scales sub-linearly with the distance if the network is in the supercritical phase.

We consider large-scale networks with n nodes, out of which k are in possession, (e.g., have sensed or collected in some other way) k information packets. In the scenarios in which network nodes are vulnerable because of, for example, limited energy or a hostile environment, it is desirable to disseminate the acquired information throughout the network so that each of the n nodes stores one (possibly coded) packet and the original k source packets can be recovered later in a computationally simple way from any (1 + isin)k nodes for some small isin > 0. We developed two distributed algorithms for solving this problem based on simple random walks and Fountain codes. Unlike all previously developed schemes, our solution is truly distributed, that is, nodes do not know n, k or connectivity in the network, except in their own neighborhoods, and they do not maintain any routing tables. In the first algorithm, all the sensors have the knowledge of n and k. In the second algorithm, each sensor estimates these parameters through the random walk dissemination. We present analysis of the communication/transmission and encoding/decoding complexity of these two algorithms, and provide extensive simulation results as well.

We consider a distributed storage problem in a large-scale wireless sensor network with n nodes among which k acquire (sense) independent data. The goal is to disseminate the acquired information throughout the network so that each of the n sensors stores one possibly coded packet and the original k data packets can be recovered later in a computationally simple way from any (1 + ∈)k of nodes for some small ∈ ≫ 0. We propose two Raptor codes based distributed storage algorithms for solving this problem. In the first algorithm, all the sensors have the knowledge of n and k. In the second one, we assume that no sensor has such global information.

In wireless networks, node mobility may be exploited to assist in information dissemination over time. We analyze the latency for information dissemination in large-scale mobile wireless networks. To study this problem, we map a network of mobile nodes to a network of stationary nodes with dynamic links. We then use results from percolation theory to show that under a constrained i.i.d. mobility model, the scaling behavior of the latency falls into two regimes. When the network is not percolated (subcritical), the latency scales linearly with the initial Euclidean distance between the sender and the receiver; when the network is percolated (supercritical), the latency scales sub-linearly with the distance.

Percolation theory has become a useful tool for the analysis of large-scale wireless networks. We investigate the fundamental problem of characterizing the critical density lambdac(d)for d-dimensional Poisson random geometric graphs in continuum percolation theory. By using a probabilistic analysis which incorporates the clustering effect in random geometric graphs, we develop a new class of analytical lower bounds for the critical density lambdac(d). These analytical lower bounds are the tightest known to date, and reveal a deep underlying relationship between clustering effects and percolation phenomena.

In battery-constrained wireless sensor networks, it is important to employ effective energy management while maintaining some level of network connectivity. Viewing this problem from a percolation-based connectivity perspective, we propose a fully distributed energy management algorithm for large-scale wireless sensor networks. This algorithm allows each sensor to probabilistically schedule its own activity based on its node degree. This mechanism is modelled by a degree-dependent dynamic site percolation process on random geometric graphs. We specify the conditions under which the resulting network is guaranteed to be percolated at all the time. We further study the delay performance of the proposed energy management algorithm by modelling the problem as a degree-dependent first passage percolation process on random geometric graphs.

In IEEE 802.11 WLANs, there are two types of access modes which perform very differently under different channel or traffic conditions: the basic access mode CSMA/CA and and an RTS/CTS based mechanisms. In seeking to balance between the use of the different modes, the contribution of this paper is threefold. Firstly, we present our development of a new method capable of calculating the access delay via an analytical model. Secondly, we report a comparison of the throughput and delay performance of basic and RTS/CTS mechanisms and the analysis of which mode should be chosen to obtain better throughput and delay performance under different conditions. Finally, we propose an adaptive RTS/CST mechanism that adjusts the RTS threshold adaptively in order to achieve optimal performance in a centralized or distributed way. Numerical results, and simulation results obtained using NS2 are provided.

We first present our recent work on the non-saturation modelling and performance analysis for IEEE 802.11 distributed coordination function. We extended the previous work and developed a model capable of reflecting the non-saturation condition. Employing the model, by introducing a concept of equivalent competing entity, we designed a measurement-aided model-based call admission control scheme for IEEE 802.11e WLANs contention-based channel access - enhanced distributed coordination function (EDCF) to provide guaranteed throughput quality of service (QoS) in a statistical sense. Extensive numerical and simulation results are provided.

The new standard IEEE 802.11e is specified to support quality-of-service in wireless local area networks. A comprehensive study of the performance of enhanced distributed channel access (EDCA), the fundamental medium access control mechanism in IEEE 802.11e, is reported in this paper. We present our development of an analytical model, in which most new features of the EDCA such as virtual collision, different arbitration interframe space (AIFS), and different contention window are taken into account. Based on the model, we analyze the throughput performance of differentiated service traffic and propose a recursive method capable of calculating the mean access delay. Service differentiation functionality and effectiveness of the EDCA are investigated through extensive numerical and simulation results. The model and the analysis provide an in-depth understanding and insights into the protocol and the effects of different parameters on the performance.