Catálogo de publicaciones - libros

Data broadcast has become a promising solution for information dissemination in the wireless environment due to the limited bandwidth of the channels and the power constraints of the portable devices. In this paper, a restricted dynamic programming approach which generates broadcast programs is proposed to partition data items over multiple channels near optimally. In our approach, a prediction function of the optimal average expected delay, in terms of the number of channels, the summation of the access frequencies of data items, and the ratio of the data items, is developed by employing curve fitting. Applying this function, we can find a cut point, which may be very close to the optimal cut. Thus, the search space in dynamic programming can be restricted to the interval around the found cut point. Therefore, our approach only takes (log) time, where is number of data items and is the number of broadcast channels. Simulation results show that the solution obtained by our proposed algorithm is in fact very close to optimal one.

Minimizing energy consumption of network operations remain a major concern in wireless sensor networks due to the limited energy capacity embedded in sensor nodes. Clustering has been proposed as a potential solution to address this issue, some nodes being responsible for the data gathering of nodes located in their vicinity. However, in order to avoid inter-cluster interference, neighboring clusters must acquire different frequencies. As the specific constraints of wireless sensor networks favor a distributed approach, we analyze modified versions of distributed backtracking, distributed weak commitment and randomized algorithms with a focus on energy consumption. In this context, we find that a heuristic may achieve better results than backtracking-based algorithms.

IP technology will play a key role in beyond 3G systems, which face the great challenge of integration in order to provide seamless service to users anywhere and anytime. Apart from its natural role as a unifier, IP also comprises the main drive for network evolution towards all-IP network infrastructures. In this regard, we exploit IP as an enabler for the evolution of the UMTS packet-switched core network, eliminating its duality at user and transport level. We focus on mobility management in the core network, which is handled by pure IP mechanisms (Mobile IPv6, MIPv6), and on the support of fast handoff across UMTS access networks by means of the IETF’s Fast MIPv6 proposal. Emphasis is put on identifying the proper interaction points between the Fast MIPv6 operation and the UMTS-specific Serving Radio Network Subsystem (SRNS) relocation procedure in order to provide a seamless handoff service to the user while not compromising the network’s performance and scalability.

The addition of storage capacity in network nodes for the caching or replication of popular data objects results in reduced end-user delay, reduced network traffic, and improved scalability. The problem of allocating an available storage budget to the nodes of a hierarchical content distribution system is formulated; optimal algorithms, as well as fast/efficient heuristics, are developed for its solution. An innovative aspect of the presented approach is that it combines all relevant subproblems, concerning node locations, node sizes, and object placement, and solves them jointly in a single optimization step. The developed algorithms may be utilized in content distribution networks that employ either replication or caching/replacement.

With the proliferation of novel paradigms in distributed systems, including service-oriented computing, ubiquitous computing or self-organizing systems, an efficient distributed management system needs to work effectively even in face of incomplete management information, uncertain situations and dynamic changes. In this paper, Bayesian networks are proposed to model dependencies between managed objects in distributed systems management. Based on probabilistic backward inference mechanisms the so-called Strongest Dependency Route (SDR) algorithm is used to compute the set of most probable faults that may have caused an error or failure.

Active networks provide an ideal support for the incorporation of intelligent behaviors into networks. This ones make possible to introduce a more general functionality, that supports the dynamic modification of the behavior of switching networks. By means of this approach, users can dynamically insert code into nodes, thus adding new capabilities.

In this paper we propose a protocol to improve network services, implemented from the point of view of active networks. The specific purpose of the protocol is to reduce the network traffic, thus increasing the ratio of the volume of transmitted data to the number of frames that pass through the network. More precisely, we intend to merge data coming from the same node or from different nodes in intermediate points of the path toward the receiver, so that they arrive as closely grouped as possible. This would reduce the congestion in networks that support a great volume of traffic, like those that are used in industrial environments, which can also suffer from a low bandwidth.

For the implementation of this protocol we use ANTS (), a freeware tools for the construction of active networks developed by MIT(). ANTS makes use of one of the most innovative and daring approaches for injecting programs into active nodes, by means of a mobile code called .

In this paper, we propose an extension to DiffServ QoS architecture in order to enhance its performance and its flexibility when used in MANETs and its adaptation to the characteristics of these networks. Then we present a formal model of our proposed extension using stochastic process algebras in order to verify the correctness and the efficiency of the proposed extension.

We propose a novel cross-layer analytic approach to performance evaluation of delay/loss sensitive IP-based applications running over the wireless channels. We firstly extend the small-scale propagation model representing the received signal strength to IP layer using the cross-layer mapping. Then, we replace the resulting IP layer wireless channel model by an artificial equivalent arrival process using the error/arrival mapping. To get performance parameters of interest we use this process together with arrival process modeling the traffic source as an input to the queuing system with deterministic service time, limited number of waiting position and non-preemptive priority discipline representing the IP packet service process of the wireless channel.

Packet-based optical ring becomes the standard access medium in metropolitan networks. Its performance depends mainly on how optical resource sharing, among different competing access nodes, takes place. This network architecture has mostly been explored in regard to synchronous transmission. However, in the present paper, we focus on the performance of asynchronous transmission-based metropolitan networks with variable packet sizes. Analytical models are presented in an attempt to provide explicit formulas that express the mean access delay of each node of the bus-based optical access network. In addition, we prove that in such a network, fairness problems are likely to arise between upstream and downstream nodes sharing a common data channel. Furthermore, we show that sharing the available bandwidth fairly and arbitrarily between access nodes, as in slotted WDM rings, does not resolve the fairness problem in asynchronous system. A basic rule, in order to achieve fairness, consists in avoiding random division of the available bandwidth caused by the arbitrary transmission of the upstream nodes.

DiffServ model is known to be the most used model to handle QoS over IP networks. Moreover this model has the advantage to be appropriate for use on large network contrary to the IntServ model that suffers from scalability problem. However, the DiffServ model has two major difficulties: routers’ configuration and resources’ allocation problems. In this paper, we introduce a new approach based on customers’ Service Level Agreements (SLA) declaration. The resource allocation is done by a federal entity called Bandwidth Broker implemented using Web-Services. Our proposal avoids the use of signaling protocol between the Bandwidth Broker and the core routers when establishing a new flow. Thus, core routers do not have the responsibility to store the customers’ traffics information and therefore, we respect the DiffServ model philosophy. Our tool provides the admission control and resource allocation management using overbooking techniques which guarantees the performances of priority traffics.