Catálogo de publicaciones - libros

The network performance obtained from the active probe packets is not equal to the performance experienced by users. We can obtain more exact result by using the characteristics of packets gained by passive measuring to calibrate the result of active measuring. The method of combining passive and active approaches has some advantages such as protocol-independent, negligible extra traffic, convenience and being able to estimate individual user performance. Considering the number of user data packets arriving between probe packets and the latency alteration of neighborhood probe packets, we propose the Pcoam (Passive Calibration of Active Measurement) method. It could reflect the actual network status more exactly, especially in the case of network congestion and packet loss, which has been validated by simulation.

Studies on Internet topology are based on measurement data. There are three types of Internet topology measurements using different sampling methodologies. This paper compares all of the three measurements together by examining their topological properties, in particular the recently introduced structural metric of rich-club connectivity. Numerical results show that, although having similar degree distribution, the topological discrepancies among the three Internet measurements are significant. This work provides a “graph-centred” analysis on the limitations of each sampling methodologies that are responsible for the measurement deficiencies.

To quantify the statistical dynamics of the BGP, we analyze the temporal and spatial correlation of macroscopic BGP message flows obtained by passive measurement. We show that the time series for the number of announcement and withdrawal messages has little correlation in time, unlike the statistical behavior of traffic volumes. This indicates that there is little possibility of a cascading failure, in which a failure causes following failures, and that the occurrence of burst of BGP messages has a Poisson nature. We also point out that there is space correlation with the delay between the flows for the different measurement points. Namely, even from macroscopic and passive measurement, we show that the propagation delay of routing information from one measurement point to another point can be statistically estimated.

Overlay multicast has been proposed as an alternative scheme to provide one-to-many or many-to-many data delivery on Internet. However, since data delivery is entirely dependent on replications on each group member, if one member cannot receive a data packet, none of its children can receive that packet. Furthermore, the higher the member’s level is, the more nodes cannot also receive data packet. In this paper, we give a detailed framework to enhance packet delivery ratio in overlay multicast. Unlike previous efforts based on duplicated forwarding, our scheme builds another type of overlay data delivery tree (DDT), which is adaptively reconstructed based on the number of group member’s measured packet delivery ratio while guaranteeing end-to-end delay bound. Through practical simulation results, we analyzed packet delivery ratio, control overhead, and end-to-end delay.

This paper proposes a rerouting scheme that can be applied to the restoration of working Label Switched Paths (LSPs) and preprovisioned backup LSPs, which consists of two subsequent algorithms. The first algorithm for the dynamic deterimination of the restoration scope (RS)increases the restoration speed by minimizing the complexity of the network topology and by maximizing the reusability of the existing working LSP. The second newly proposed concept of RS extension minimizes the probability of restoration failure by dynamically widening the restoration scope until the RS is equal to the whole network topology. Through simulation, we evaluate the performance of our restoration scheme and the existing protection schemes in terms of the restoration speed, packet loss, network resource utilization, and resource reusability of the existing working LSP.

Traditionally, there is an assumption that the QoS-aware multicast routing protocol acquires QoS information via the QoS-aware unicast routing protocol. Mostly, the research focuses on about managing group dynamics and failure recovery on multicast and there is no interest in the QoS-aware unicast routing protocol for multicast. In fact, researches of the QoS-aware multicast based on QoS-aware unicast are uncommon. Hence, we focus on the QoS-aware unicast routing algorithm and apply it to multicast. In this paper, we use a novel algorithm, called QoS Restricted and Distributed Generic Shortest Path Algorithm (QRDGSPA). As it is a very simple and measured method without complex computation Hence, it can make a shortest path or QoS-aware multiple paths. For this purpose, we use a specific routing algorithm, Multicast Candidate Selection Method (MCSM), which can choose the optimal candidate path from a receive node to a candidate node. Based on integration of QRDGSPA and MCSM, we propose QoS-aware and group Density-aware Multicast Routing Protocol (QDMRP).

In this paper, we attempt to solve the problem of constructing a minimum cost multicast tree with the consideration of dynamic user membership. Unlike the other minimum cost multicast tree algorithms, this problem consists of one multicast group of fixed members and each destination member is dynamic and has a probability of being active as which was gathered by observation over some period of time. With the omission of node join/leave handling, this model is suitable for prediction and planning purpose than for online maintenance of multicast trees. We formally model this problem as an optimization problem and apply the Lagrangean relaxation method and the subgradient method to solve the problem. Computational experiments are performed on regular networks and random networks. According to the experiment results, the Lagrangean based heuristic can achieve up to 37.69% improvement compared to the simple heuristic.

In wireless sensor networks, the sensors collect data and deliver it to a sink node. Most of the existing proposals deal with the traffic flow problem to deliver data to the sink node in an energy-efficient manner. In this paper, we extend this problem into a multi-sink case. To maximize network lifetime and to ensure fairness, we propose (i) how to position multiple sink nodes in a sensor network and (ii) how to route traffic flow from all of the sensors to these multiple sink nodes. Both of the problems are formulated by the linear programming model to find optimal locations of the multiple sink nodes and the optimal traffic flow rate of routing paths in wireless sensor networks. The improved lifetime and fairness of our scheme are compared with those of the multi-sink aware minimum depth tree scheme.

Since the operations of sensors in a wireless sensor network mainly rely on battery power, power consumption becomes an important issue. In this paper, we will consider the problem of searching for multiple paths between multiple source sensors and the sink such that any sensor in a path from a source sensor to the sink does not run out of its power during the transmission of packets. The problem has been proved to be NP-complete. Based on the principle of genetic algorithms, in this paper, we will design an efficient heuristic algorithm for it. Computer simulations verify that the suboptimal solutions generated by our genetic algorithm are very close to the optimal ones.

This paper proposes E-MAC, a contention-based energy-efficient Medium Access Control (MAC) Protocol for wireless sensor networks. Energy efficiency is primary goal in wireless sensor networks. Existing MAC protocols for sensor networks attempt to solve energy consumption problem caused by idle listening using an active/sleep duty cycle. Since there are various traffic conditions, however, they may not always provide improvements in energy consumption. We propose a MAC protocol algorithm that stores data in a buffer and transmits data when the buffer exceeds a threshold value so that energy efficiency is always guaranteed for any network traffic conditions. Analytical results show that our proposed algorithm has significant improvements in energy consumption compared to the existing MAC protocols for sensor networks.