Catálogo de publicaciones - libros

A queueing model for the system level performance evaluation of mobile cellular networks considering both resource insufficiency and wireless channel unreliability is proposed and mathematically analyzed. The proposed mathematical approach is based on the use of simple call interruption processes to model the effect of wireless channel unreliability. More importantly, this paper develops a system level teletraffic model for the performance evaluation considering that channel holding times for new and handed off calls are general (but not necessary identically) distributed random variables. Additionally, an approximated one-dimensional recursive approach based on the well known Kauffman-Roberts formula is proposed for the case when the channel holding time variables can be adequately characterized by Hyper-Exponential distributions. Also, the case when the channel holding time variables are Mixed-Erlang distributed, a distribution having universal approximation capability, is analyzed. Thus, our teletraffic model allows obtaining more general, realistic, and easily computable analytical results.

We propose a novel performance model for the HSDPA in presence of circuit-switched dedicated channels. The model consists of two parts: An HSDPA bandwidth model which considers the SIR distribution according to the multi-path model and the number of available channelization codes, and an analytical capacity model which integrates HSDPA and dedicated channels under assumption of adaptive resource allocation for the HSDPA. Additionally, the model considers the impact of location dependent bandwidths on the spatial user distribution. The accuracy of the model is demonstrated for an example scenario.

We focus in this paper on passive traffic measurement techniques that collect traces of TCP packets and analyze them to derive, for example, round-trip times or aggregate metrics such as average throughput. The seminal work of Zhang [1] has shown that for more than 50% of the TCP connections observed, it is not the network bandwidth that limits the throughput but rather the application or mechanisms such as TCP slow start or too small a receiver window. Certain types of analysis of the network characteristics are meaningful only when performed on TCP traffic that experiences minimal interference by the application. To eliminate such interference, we propose a generic method that partitions the packets of a TCP connection in bulk data transfer and in application limited periods: The packets of a bulk data transfer period (BTP) experience minimal interference from the application, while the packets of an application limited period (ALP) experience interference from the application that prevents TCP from fully utilizing the network resources because the application does not produce data fast enough. As a proof of concept, we apply our algorithm to public Internet traffic traces and show that unless the effects of the application are filtered out, studying the end-to-end path and traffic characteristics from a network point of view can produce biased results.

We analyze in this paper eDonkey traffic on the basis of measurements from the France Telecom IP collect network. Specifically, we investigate the structure of the eDonkey network by analyzing addresses and port numbers involved in eDonkey transactions. This analysis allows us to understand the dynamics of eDonkey traffic in a subnetwork of the whole eDonkey network. In particular, we stress the fact that eDonkey gives rise to a huge amount of signaling, identified as small TCP data transfers. To qualitatively explain the observed phenomena, we develop a very simple qualitative model to understand the dynamics of an eDonkey community sharing a given object (a chunk or a file). This model is intended to capture the transient behavior of a network relying on eDonkey principles when sharing an object. This simple model allows us to identify two regimes depending on the congestion level of servers (expansion and collapse).

The purpose of the paper is two-fold: first, characterize the Internet data traffic at high levels, i.e., flows and sessions, with the principal aim of comparing the salient features shown by different types of applications, namely P2P on one hand and classical Internet applications in the other. Second, among the observed traffic characteristics, we particularly focus on the way bandwidth is actually shared by flows and users by developing an empirical bandwidth sharing model. We infer from it some trends shown by the different classes of traffic in terms of elasticity, i.e., their potential reaction against congestion.

We investigate statistical properties of the traffic especially for ADSL broadband access platforms, which have been widely deployed in recent years. Measurement traces of aggregated traffic are evaluated on multiple time scales and show an unexpected smooth profile with less relevance of long range correlation than experienced for traffic from Ethernet LANs.

A reason for the different characteristics lies in the shift to an increasing population of residential users generating most traffic on IP platforms via ADSL access. In addition, data transfer protocols of peer-to-peer networks strengthen the smoothing effect observed in current IP traffic profiles.

While the problem of optimal cell-site partitioning has been primarily studied from the perspective of scarce radio resources in the access network, recent field measurements have shown significant impact of paging load on the core signaling subsystem capacity. In this paper, we revisit the problem of optimal zone partitioning, identifying factors affecting partitioning cost at the core switch. We develop a general integer programming formulation for joint optimization of two key issues: (i) assignment of cells to a base station controller and controllers to the core signaling subsystem, (ii) optimal partitioning of zones at the access and network levels. Given the exponential nature of the location planning problem, we develop a genetic algorithm based approach for solving the general zone partitioning and configuration problem, both for incumbent and greenfield networks. Our results demonstrate significant cost and performance benefits at the network level for next generation converged services.

Spatial and temporal load variations, e.g. flash overloads and traffic hot spots that persist for minutes to hours, are intrinsic features of wireless networks, and give rise to potentially huge performance repercussions. Dynamic load balancing strategies provide a natural mechanism for dealing with load fluctuations and alleviating the performance impact. In the present paper we propose a distributed shadow-price-based approach to dynamic load balancing in wireless data networks. We examine two related problem versions: (i) minimizing a convex function of the transmitter loads for given user throughput requirements; and (ii) maximizing a concave function of the user throughputs subject to constraints on the transmitter loads. As conceptual counterparts, these two formulations turn out to be amenable to a common primal-dual decomposition framework. Numerical experiments show that dynamic load balancing yields significant performance gains in terms of user throughputs and delays, even in scenarios where the long-term loads are perfectly balanced.

In this paper, we propose a model for the global design problem of wavelength division multiplexing (WDM) networks including the traffic grooming. This problem consists in finding the number of fibers between each pair of node, selecting the configuration of each node, choosing the set of lightpaths (i.e., the virtual topology), routing these lightpaths over the physical topology and finally, grooming and routing the traffic over the lightpaths. Since this problem is NP-hard, we propose heuristic algorithms and a tabu search metaheuristic algorithm to find good solutions for real-size instances rapidly.

Recent advances in telecommunication networks have allowed WDM to emerge as a viable solution to the ever-increasing demands of the Internet. In a wavelength-routed optical network, traffic is transported over lightpaths, which exclusively occupy an entire wavelength on each hop of the source-destination path. Because these networks carry large amounts of traffic, alternate routing methods are designed in order to allow traffic to be properly re-routed from source to destination in the event of certain events, such as link blocking or failure. In this paper, we consider a tandem traffic-groomed optical network, modeled as a multi-level overflow system, where each level represents a wavelength between adjacent nodes. The queueing network is analyzed using a combination of methods. As will be shown, the decomposition method provides a good approximate analysis of large overflow systems supporting traffic from multiple sources.