Catálogo de publicaciones - libros

We consider jamming in wireless networks with transmission cost for both transmitter and jammer. We use the framework of non-zero-sum games. In particular, we prove the existence and uniqueness of Nash equilibrium. It turns out that it is possible to provide analytical expressions for the equilibrium strategies. These expressions is a generalization of the standard water-filling. In fact, since we take into account the cost of transmission, we obtain even a generalization of the water-filling in the case of one player game. The present framework allows us to study both water-filling in time and water-filling in frequency. By means of numerical examples we study an important particular case of jamming of the OFDM system when the jammer is situated close to the base station.

The Tit-for-Tat strategy implemented in BitTorrent (BT) clients is generally considered robust to selfish behaviours. The authors of [1] support this belief studying how Tit-for-Tat can affect selfish peers who are able to set their upload bandwidth. They show that there is a “good” Nash Equilibrium at which each peer uploads at the maximum rate. In this paper we consider a different game where BT clients can change the number of connections to open in order to improve their performance. We study this game using the analytical framework of network formation games [2]. In particular we characterize the set of the peers can form and how the peers can dynamically reach such configurations. We also evaluate the loss of efficiency peers experience because of their lack of coordination: we find that the loss of efficiency is in general unbounded despite the utilization of the Tit-for-Tat strategy.

We consider a collision channel, shared by a finite number of self-interested users with heterogenous throughput demands. It is assumed that each user transmits with a fixed probability at each time slot, and the transmission is successful if no other user transmits simultaneously. Each user is interested in adjusting its transmission rate so that its throughput demand is met. When throughput requirements are feasible, we show that there exist two equilibrium points where users satisfy their respective demands. In one equilibrium all users transmit at lower rates, compared to their transmission rates at the other equilibrium. This fact is meaningful in wireless systems, where lower transmission rates translate to power savings. Subsequently, we propose a distributed scheme that ensures convergence to the lower-rate equilibrium point. We also provide some lower bounds on the channel throughput that is obtained with self-interested users, both in the symmetric and non-symmetric case.

We consider the problem of selfish or competitive routing over a network with flow-dependent costs which is shared by a finite number of users, each wishing to minimize the total cost of its own flow. The Nash Equilibrium is well known to exist for this problem under mild convexity assumptions on the cost function of each user. However, uniqueness requires further conditions, either on the user cost functions or on the network topology. We briefly survey here existing results that pertain to the uniqueness issue. We further consider the mixed Nash-Wardrop problem and propose a common framework that allows a unified treatment of this problem.

Making drop decisions to enforce the max-min fair resource allocation in a network of standard TCP flows without any explicit state information is a challenging problem. Here we propose a solution to this problem by developing a suite of stateless queue management schemes that we refer to as Multi-Level Comparison with index l (MLC()). We show analytically, using a Markov chain model, that for an arbitrary network topology of standard TCP flows and queues employing MLC(), the resource allocation converges to max-min fair as increases. The analytical findings are verified experimentally using packet level 2 simulations.

Understanding the relationship between queueing delays and link utilization for general traffic conditions is an important open problem in networking research. Difficulties in understanding this relationship stem from the fact that it depends on the complex nature of arriving traffic and the problems associated with modelling such traffic. Existing AQM schemes achieve a “low delay” and “high utilization” by responding early to congestion without considering the exact relationship between delay and utilization. However, in the context of exploiting the delay/utilization tradeoff, the optimal choice of a queueing scheme’s control parameter depends on the cost associated with the relative importance of queueing delay and utilization. The optimal choice of control parameter is the one that maximizes a benefit that can be defined as the difference between utilization and cost associated with queuing delay. We present a generic algorithm Optimal Delay-Utilization control of (ODU-) that is designed with a performance goal of maximizing this benefit. Its novelty lies in fact that it maximizes the benefit in an online manner, without requiring knowledge of the traffic conditions, specific delay-utilization models, nor does it require complex parameter estimation. Moreover, other performance metrics like loss rate or jitter can be directly incorporated into the optimization framework as well. Packet level ns2 simulations are given to demonstrate the behavior of the proposed algorithm.

In this paper, we introduce two different models of FAST TCP. One is from the original FAST TCP model, and the other is from the implementation of FAST TCP. Interestingly, these two models show significantly different dynamic performance and stability characteristics. That is, while the latter model is always globally exponentially stable, the former model is faster than the latter model and possible to go unstable. Motivated from these two models, we suggest a modified congestion control algorithm. By tuning the gain of the terms caused by the difference of the two models, the modified algorithm can be made to become globally asymptotically stable and more responsive than the implementation model of FAST TCP. The stability condition of the modified algorithm is decoupled from the network parameters and does not change the equilibrium state.

We propose a new MIMD Adaptive RED and revisit two well-known adaptive algorithms for ARED, (Feng et al., A Self Configuring RED Gateway, , 1999) and (Floyd et al., Adaptive RED). We use the steady-state relation between the maximum marking (dropping) probability, max, and the average queue length, , to argue that different adaptive schemes, AIMD, MIMD, etc., can be proposed when max varies slowly. We model MIMD ARED and study stability. It is shown through simulations that the performance and robustness properties of MIMD ARED are in fact similar to those of Floyd’s ARED.

In the networking research literature, the problem of network utility optimization is often converted to the dual problem which, due to nondifferentiability, is solved with a particular subgradient technique. This technique is not an ascent scheme, hence each iteration does not necessarily improve the value of the dual function. This paper examines the performance of this computational technique in realistic mesh network settings. The traditional subgradient technique is compared to a subgradient technique that is an ascent algorithm. It is found that the traditional subgradient techniques suffer from poor performance. Specifically, for large networks, the convergence is slow. While increasing the step size improves convergence speed, due to stability problems, the step size cannot be set arbitrarily high, and suitable step sizes result in slow convergence. The traditional subgradient technique also suffers from difficulties when used online. The ascent scheme performs well in all respects, however, it is not a distributed technique.

We consider channel allocation to mitigate interference between wireless LANs. The channel allocation task is often formulated in the literature as finding a proper colouring of a single graph. We show that the interference between WLANs can be channel dependent in which case a different conflict graph is associated with each channel. Channel allocation then corresponds to a multi-graph colouring problem. This potentially has profound implications as the behaviour of many proposed colouring-based algorithms for channel allocation is unclear in a multi-graph context. We prove that a recently proposed decentralized colouring algorithm performs correctly in the multi-graph setting. We also present a new, extended version of this algorithm suited to a wide range of multi-radio architectures.