Catálogo de publicaciones - libros

The increasing use of wireless networks and the popularity of multimedia applications, leads to the need of QoS (Quality of Service) support in a mobile IP-based environment. This paper presents the framework, needed to support both micromobility and resource reservations. We present an admission control mechanism in which a mobile host can trigger reservations without performing handoff, taking advantage of link state changes caused by the handoff of other mobile hosts. We also investigate the impact of inaccurate link state information and the occurrence of simultaneous handoffs on the performance of the handoff and reservation mechanism. This impact is higher when only a small part of the mobile hosts can receive QoS service at the same time. For the simulations, we use Q-MEHROM [10]. Herein, [11] gathers the link state information and calculates the QoS tables. However, the ideas and results presented in this paper are not restricted to these protocols.

Handoff latency affects the service quality of real-time applications. In this paper we develop an analytical model to analyze the Mobile IP Fast Authentication protocol (MIFA) and compare it to Hierarchical Mobile IP (HMIP). The study compares the signaling costs of the protocols as well as the overall load for packet forwarding. Our study shows that MIFA minimizes the packet delivery cost compared to HMIP. Additionally, MIFA is more efficient when the arrival rate of the packets increases. Thus MIFA outperforms HMIP with respect to signaling cost. From the performance point of view MIFA performs similar to HMIP when the domain consists of two hierarchy levels only, and outperform HMIP otherwise. However, MIFA does not require a hierarchical network architecture as HMIP does.

High level services and data transport service provision are facing important advancements towards a more flexible business models and Internet services with the internetworking of several complementary access technologies using IP. This imposes new requirement in users’ mobile terminal, such as intelligent discovery and selection of access networks, vertical handover support and roaming. The result of such integration of networks, also known as “4th Generation Networks”, will require that transport network and user’s terminal coordinates for providing of a “persistent” transport service. This paper presents the conceptual high level description of a context-based management distributed service and profiling system based in OWL, the W3C ontology language, in order to provide this persistent transport for mobile user’s data in a heterogeneous environment of IP-based network.

File replication for uninterrupted availability is affected by the localised nature of network failures, particularly in ubiquitous, mobile environments; nearby nodes often get disconnected together, as a result of switching equipment faults, or of local wireless network unavailability – for instance, failure of a base station, or loss of network connectivity when a train enters a tunnel.

In this paper we propose replic8, a substrate for location-aware file replication, mitigating the effect of localised network failures by storing replicas at network locations selected for being far away. We demonstrate that, compared to storage of replicas at random network locations, replic8 achieves high data availability, and requires lower numbers of replicas to maintain that.

This paper presents a simple and accurate analytical model of TCP Reno throughput as a function of loss rate, average round trip time and receiver window size based on PFTK-model. The presented model refines previous work by careful examination of fast retransmit/fast recovery dynamics in the presence of correlated losses and taking into consideration slow start phase after timeout. The accuracy of the proposed model is validated against simulation results and compared with those of PFTK-model. Simulation results show that our model gives a more accurate estimation of TCP Reno throughput in the presence of correlated losses than PFTK-model.

The diversity of the networks (wired/wireless) prefers a TCP solution robust across a wide range of networks rather than fine-tuned for a particular one at the cost of another. TCP parallelization uses multiple virtual TCP connections to transfer data for an application process and opens a way to improve TCP performance across a wide range of environments – high bandwidth-delay product (BDP), wireless as well as conventional networks. In particular, it can significantly benefit the emerging high-speed wireless networks. Despite its potential to work well over a wide range of networks, it is not fully understood how TCP parallelization performs when experiencing various packet losses in the heterogeneous environment. This paper examines the current TCP parallelization related methods under various packet losses and shows how to improve the performance of TCP parallelization.

In this paper, we investigate extensively the joint network dynamics with different AIMD window-adjustment parameters on end-hosts, and different queue management schemes (i.e. DropTail vs. RED) in routers. We reveal that with DropTail buffer, although smooth TCPs causes less queuing-delay jitter, its average queuing delay is significantly higher than that of responsive TCPs. The direct implication of this discovery is that when mobile users of media-streaming and short messages share a bottleneck link, the energy consumption for sending short messages can increase severely if media-streaming users adopt smooth TCPs. With RED, on the other hand, smooth TCPs not only lead to smaller queue oscillation, the average/max queue length is smaller as well.

In this paper, we present a novel congestion control algorithm for the Transmission Control Protocol (TCP) for the future Internet. Our assumption of future Internet is that, with the increasing quality of service (QoS) requirements, per-flow packet scheduling (per-flow here refers to per TCP or UDP connection) will replace the current first-come-first-serve algorithm used in routers. Based on the assumption, we design a new congestion control algorithm. In our TCP-CC algorithm, each connection adjusts the size of the congestion window according to the size of its packet queue at the bottleneck router. Thus, the queue size for each connection at the bottleneck router is within a controlled range. We show that congestion loss is effectively reduced compared to the current TCP congestion algorithm.

The work describes the performance of two congestion control algorithms: AIMD and -AIMD. The first is the default mechanism of TCP; the second is a proposed congestion control algorithm that improves fairness of AIMD. We consider asynchronous networks where TCP flows have different propagation delays, a portion of their link is wireless, and they compete for resources over a single bottleneck link. We show that -AIMD improves the performance of flows that have long propagation delay and the fairness of the network. In wireless links -AIMD outperforms AIMD and the cost of lost packets (or wasted energy) is the same as that of AIMD.

Given that applications and services for evolving mobile ad hoc networks (MANETs) have diverse quality of service requirements in a similar fashion to fixed networks, this paper proposes a proportional service differentiation (PSD) model. This model is highly scalable and simple to adopt in MANETs because it does not require explicit admission control or maintenance of state information in any intermediate node. It relies instead on localized scheduling and buffer management to achieve a desired global objective. Motivated by this aspect of the PSD model, we propose to combine it with a location-based forwarding strategy as a way to facilitate cross-layer optimization. This association is performed with a view to improve end-to-end service differentiation, although no other explicit mechanisms are used to achieve end-to-end guarantees. This model takes also into consideration the time-varying nature of available bandwidth in MANETs, and tries to calculate it dynamically. Simulation results confirm the per-hop performance improvement.