Catálogo de publicaciones - libros

We consider a fluid queue in discrete time with random service rate. Such a queue has been used in several recent studies on wireless networks where the packets can be arbitrarily fragmented. We provide conditions on finiteness of moments of stationary delay, its Laplace-Stieltjes transform and various approximations under heavy traffic. Results are extended to the case where the wireless link can transmit in only a few slots during a frame.

In this paper, we analyse a finite sized discrete time priority buffer with two types of packet arrivals, referred to as class 1 and class 2 packets. Packets of both classes arrive according to a discrete batch Markovian arrival process, taking into account the correlated nature of arrival processes in heterogeneous telecommunication networks. Packets of class 1 are assumed to have both space priority and transmission priority over packets of class 2. In particular, the partial buffer sharing acceptance policy is adopted as space priority mechanism. Using a matrix analytic approach, the buffer content is analysed and through some numerical examples, the impact of the priority scheduling, the threshold and the correlation in the arrival process is demonstrated.

Measurement studies have shown that the probabilities of file possession and the arrival rates of file requests have an important impact on the file downloading times in p2p file sharing systems resulting in unfair performances. A model has been proposed to capture this phenomenon. This model is used to derive the qualitative impact of system parameters on performances. However to obtain results for realistic p2p networks one is confronted with the large number of files in these systems. To show the influence of the number of files shared in the system we present results obtained by simulations in the case of small to medium size systems.

We consider a BitTorrent-like file sharing system, where the peers interested in downloading a large file join an overlay network. The seed node possessing the file stays in the system, whereas all other peers are non-altruistic in the sense that they leave the system as soon as they have downloaded the whole file. We consider a flash crowd scenario, where the peers join the overlay simultaneously. We show that the chunk selection algorithm is critical, propose an analytic approach to the process, and find that the encounters can be restricted to neighbours in a Chord overlay without losing much in performance.

The prevalent mechanism to avoid congestion in IP networks is the control of the sending rate with TCP. Dynamic routing strategies at the IP layer are not deployed because of problems like route oscillations and out-of-order packet deliveries.

With the adoption of P2P technology, routing is done also in these overlay networks. With multi-source download protocols peers upload and download to/from other peers in parallel.

Based on congestion pricing for IP networks this paper proposes a rate control algorithm for P2P over IP networks. A peer adopts the functionality of TCP and extends the congestion window mechanism with information from the overlay network. Thus, a sending peer is able to shift traffic from a congested route to an uncongested one. This change in the rate allocation will be balanced by other peers in the overlay. Hence, the receiving peers experience no degradation of their total download rate.

In structured peer-to-peer (P2P) networks participating peers can join or leave the system at arbitrary times, a process which is known as churn. Many recent studies revealed that churn is one of the main problems faced by any Distributed Hash Table (DHT). In this paper we discuss different possibilities of how to estimate the current churn rate in the system. In particular, we show how to obtain a robust estimate which is independent of the implementation details of the DHT. We also investigate the trade-offs between accuracy, overhead, and responsiveness to changes.

This paper evaluates the performance of two schemes for recovering lost data in a peer-to-peer (P2P) storage systems. The first scheme is centralized and relies on a server that recovers multiple losses at once, whereas the second one is distributed. By representing the state of each scheme by an absorbing Markov chain, we are able to compute their performance in terms of the delivered data lifetime and data availability. Numerical computations are provided to better illustrate the impact of each system parameter on the performance. Depending on the context considered, we provide guidelines on how to tune the system parameters in order to provide a desired data lifetime.

Traffic matrices constitute essential inputs in a wide variety of network planning and management functions as they provide the traffic volumes that flow between the node pairs in a network. In operational IP networks, it is desirable that traffic matrix (TM) estimation relies on information that is directly obtainable from SNMP system measurements i.e., link counts data. Existing approaches for TM estimation based on link counts have been shown to have limited accuracy and cannot be generally applied to practical IP networks. In this paper, we propose a new method for TM estimation which makes more accurate assumptions about the traffic characteristics of the flows between node pairs and, specifically, a Markovian Arrival Process of order two (MAP-2) has been applied for this purpose. The presented evaluation study shows the ability of the method to accurately capture the correlation and burstiness statistics of real IP flows and, therefore, can be successfully applied in IP network management functions.

Network traffic measurements can provide essential data for network research and operation. While Internet traffic has been heavily studied for several years, there are new characteristics of traffic having not been understood well brought by new applications for example P2P. It is difficult to get these traffic metrics due to the difficulty to measurement traffic on line for high speed link and to identify new applications using dynamic ports. In this paper, we present a broad overview of Internet traffic of an operated OC-48 export link of a metro area network from a carrier with the method of measurement on-line. The traffic behaves a daily characteristic well and the traffic data of whole day from data link layer to application layer is presented. We find the characteristics of traffic have changed greatly from previous measurements. Also, we explain the reasons bringing out these changes. Our goal is to provide the first hand of traffic data that is helpful for people to understand the change of traffic with new applications.

Under the assumption that packets are sufficiently interleaved and the sampling rate is small, we show in this paper that those characteristics of flows like the number of packets, volume, etc. obtained through deterministic 1-out-of- packet sampling is equivalent to random packet sampling with rate  = 1/. In particular, under mild assumptions, the tail distribution of the total number of packets in a given flow can be estimated from the distribution of the number of sampled packets. Explicit theoretical bounds are then derived by using technical tools relying on bounds of Poisson approximation (Le Cam’s Inequality) and refinements of the central limit theorem (Berry-Essen bounds). Experimental results from an ADSL traffic trace show good agreement with the theoretical results established in this paper.