A Selective Dropping Congestion Control (SDCC) algorithm is proposed in this work, for wireless multimedia sensor networks based on scalable multimedia encoding. SDCC, which has been designed by means of linear discrete time control theory arguments, can be applied in a fully distributed way. It periodically evaluates at each node the amount of data to drop, in order to control the level of the transmission queue, and then selects the packets to discard based on their relevance to the image quality. SDCC effectiveness has been evaluated with reference to a video-surveillance problem in several scenarios, using the Castalia simulator. Results have shown that SDCC is able to significantly improve the network performance in terms of image delivery ratio and delay.

Long-term evolution represents an emerging technology that promises a broadband and ubiquitous Internet access. But several aspects have to be considered for providing effective multimedia services to mobile users. In particular, in this work, we consider the design of a quality-of-service (QoS) aware packet scheduler for real-time downlink communications. To this aim, a novel two-level scheduling algorithm is conceived. The upper level exploits an innovative approach based on discrete-time linear control theory. Instead, at the lower level, a proportional fair scheduler has been properly tailored to our purposes. The performance and the complexity of the proposed scheme have been evaluated both theoretically and by using simulations. A comparison with recently proposed scheduling strategies has been also presented, considering several network conditions and real-time multimedia flows. Particular attention has been devoted to the evaluation of the quality-of-experience (QoE) provided to end users. Results have clearly shown that the proposed approach is able to greatly outperform the existing ones especially in the presence of real-time video flows.

The importance of a Variable Bit Rate (VBR) video transmission on UMTS networks is increasing in time. The bursty nature of VBR traffic complicates the design of efficient mechanisms for video retrieval, transport, and provisioning to achieve a high bandwidth utilization and reduce the negative effects of bandwidth fluctuations in wireless channels. To this aim, several scheduling algorithms can be successfully implemented. They regulate data transmission to reduce the rate variability peculiar of VBR streams. At client side, scheduled data are temporarily stored in the client buffer before being decoded on the terminal. In this work, a novel scheduling algorithm, the Scheduling Algorithm for Interactive Video (SAIV) is presented and analyzed. It is an algorithm thought for VBR stream transmission in UMTS networks that takes into account the user interactivity. Scheduling is performed “online”, over relatively small video segments to reduce delays. SAIV dynamically varies the sampling frequency of the Real Time Control Protocol (RTCP) feedbacks that carry information on the client buffer status. The sampling frequency is modulated according to the difference between the calculated buffer fill level at server side and the real buffer fill level at client side. The latter is exploited to reschedule data with the updated information. Numerical results testify the SAIV effectiveness compared to the classical SLWIN online algorithm already known by literature, in some simulation scenarios of real interest.

Abstract The 802.15.3 standard is a reference point for high data‐rate wireless personal area networks (WPANs), being able to support also multimedia applications with specific quality of service (QoS) requirements. In a 802.15.3 piconet, a single piconet co‐ordinator (PNC) distributes channel time allocations (CTAs) to devices (DEVs) associated to it, in order to provide the expected QoS. The 802.15.3 standard does not specify how this allocation should be done. To bridge this gap, we propose a dynamic CTA scheduling scheme that provides bounded average delays to multimedia applications. The algorithm is based on a control theoretic approach and acts in two steps: (1) each single DEV computes its transmission needs by applying a control algorithm that targets empty transmission queues; (2) the PNC collects requests of DEVs and allocates CTAs by taking into account the capacity constraints of the wireless channel. The main properties and the tuning rules of the algorithm have been theoretically investigated. Moreover, in order to provide a comprehensive performance evaluation of the proposed approach, realistic scenarios with video, voice, and FTP flows have been simulated using ns‐2. Results confirm that our scheduler is able to satisfy the expected delay bounds in a wide range of traffic loads. Copyright © 2009 John Wiley & Sons, Ltd.

A wireless networked control system exploits a shared wireless network to logically link its sub-systems. Due to the peculiarities of the wireless channel, it is important to understand in which measure the quality of service provided by the communication infrastructure affects the quality of control. The aim of this work is to analyze the performance of a wireless networked control system based on the IEEE 802.11 MAC. For that purpose, a model of such a system has been built using the ns-2 simulator and the impact of MAC parameters on the system dynamics has been evaluated. Simulation results have highlighted strengths and limitations of this kind of wireless networked control system.

In Internet multimedia streaming, a raw quality base layer is adaptively enriched with one or more enhancement layers to fully use the available bandwidth. In this work, using a control theoretic approach, we design and implement an innovative scheduling strategy for properly distributing the network available bandwidth among base and enhancement layers, in order to avoid playout interruptions due to base layer underflows at the decoder side. The proposed solution has been casted in a H.264/AVC SVC video streaming architecture. Experimental results, collected using a single bottleneck testbed and three different benchmarking video sequences, have demonstrated that the proposed scheduling strategy is able to avoid or strongly limit the number of playout interruptions under moderate or high traffic conditions, respectively.

The recent interest for wireless networked control systems has boosted new research activities on real-time 802.11-based communications. In fact, due to their versatility, wireless LANs can be exploited as infrastructure in this kind of control systems, but changes to channel access techniques are required. To this aim, we propose a novel experimental implementation of a TDMA scheme over 802.11 MAC, based on the hard real-time networking framework RTnet and on the Xenomai real-time task scheduler. Preliminary results have shown that, using the proposed approach, transmission delays smaller than 10 ms and packet loss rates smaller than 0.5% can be achieved.

CF-MAC is an energy efficient collision-free MAC protocol for wireless ad-hoc networks, based on the theory of Balanced Incomplete Block Design (BIBD). It has been originally designed considering energy saving as primary aim. In this paper, using computer simulations, its ability to handle voice calls has been evaluated. Specifically, the Mean Opinion Score and the Cumulative Distribution Function of the one way packet delay have been analyzed, for various protocol parameters and network scenarios. Moreover, a comparison between CF-MAC and classical 802.11 DCF access function has been proposed. Simulation results have shown that by properly selecting protocol parameters, CF-MAC can effectively support VoIP calls.

In order to provide service differentiation in 802.11 wireless local area networks, the 802.11e working group has recently proposed the hybrid coordination function (HCF) along with a HCF controlled channel access (HCCA). However, the 802.11e proposal does not specify an effective bandwidth allocation algorithm. Recently, to overcome this limitation, a control theoretic framework has been proposed in literature. Within such a framework, this paper proposes a new allocation algorithm based on a proportional-integral-derivative (PID) controller. Moreover, a performance comparison among the simple Scheduler introduced by 802.11e standard and three bandwidth allocations algorithms, designed using different dynamic regulators, has been carried out. Ns-2 simulation results showed that the best trade-off between packet delays and number of admitted flows is achieved by a properly tuned PID regulator.

Modeling TCP is fundamental for understanding Internet behavior. The reason is that TCP is responsible for carrying a huge quota of the Internet traffic. During last decade many analytical models have attempted to capture dynamics and steady-state behavior of standard TCP congestion control algorithms. In particular, models proposed in literature have been mainly focused on finding relationships among the throughput achieved by a TCP flow, the segment loss probability, and the round trip time (RTT) of the connection, which the flow goes through. Recently, Westwood+ TCP algorithm has been proposed to improve the performance of classic New Reno TCP, especially over paths characterized by high bandwidth-delay products. In this paper, we develop an analytic model for the throughput achieved by Westwood+ TCP congestion control algorithm when in the presence of paths with time-varying RTT. The proposed model has been validated by using the ns-2 simulator and Internet-like scenarios. Validation results have shown that this model provides relative prediction errors smaller than 10%. Moreover, it has been shown that a similar accuracy is achieved by analogous models proposed for New Reno TCP.

IEEE 802.11e Working Group has recently proposed an Enhanced Distributed Coordination Access (EDCA) scheme to improve the standard 802.11 Distributed Coordination Function (DCF) in order to provide service differentiation capabilities to WLANs. However, it has been widely demonstrated that EDCA does not provide an effective usage of the channel capacity, especially at high network loads.This paper proposes the Extended EDCA scheme as a distributed dynamic bandwidth allocation algorithm, providing delay guarantees to real-time media flows. Using the ns-2 simulator, we have shown that the proposed algorithm is able to provide a bounded delay service to multimedia flows in a wide range of network loads and frame loss ratios.

A novel strategy of measurement based admission control (MBAC) for video distribution systems has been developed, aimed at obtaining the needed quality of service (QoS), expressed in terms of loss probability. The developed algorithm is based on an autoregressive (AR) adaptive filter taking into account slow and quick time bit rate variations of video stream aggregation. A simple admission control strategy is presented, able to shift from the server to the client the decision about the admission of a new video stream. Simulation results, based on real MPEG encoded videos justify the application

Recent advances in information and communication technologies have made multimedia-on-demand services technically and economically feasible. Important aspects of such systems are the resource sharing techniques, which allow the simultaneous service of a large number of users with considerable savings in terms of network bandwidth and server resources. In this paper, we report the results of a study which analyzes batching and buffering techniques, which involves serving all video requests issued during a short interval of time with a single stream. The mathematical model, based on queueing networks, allows the evaluation of the main system performance (average and probability distribution of the number of streams, percentage reduction of resources, and so on) as a function of load and batching interval duration. Simulation experiments confirm the analytical model in the whole range of considered conditions.

The main aspect of several multimedia applications is the transmission of variable bit rate (VBR) video streams requiring high bandwidth with stringent quality of service (QoS) guarantee. In such systems, a statistical bandwidth estimation is relevant for network infrastructure planning and admission control. In this paper, an original algorithm for aggregate bandwidth estimation, needed by a given number of correlated video streams, is proposed and analyzed. The obtained analytical results are validated by simulation and compared with other numerical results already present in the literature, discussing the effectiveness of the proposed solution.

Westwood+ TCP is a sender-side only modification of the classic TCP that is based on end-to-end estimation of the bandwidth available to the connection in order to shrink adaptively the TCP congestion window and the slow start threshold after congestion. We report measurements obtained using Linux implementations of Westwood+, Westwood and Reno to FTP data over Internet connections spanning continental and intercontinental distances. In particular, we show that the bandwidth estimation algorithm employed by Westwood+ nicely tracks the available bandwidth, whereas the previous bandwidth estimation algorithm used by TCP Westwood fails to work in the real Internet due to ACK compression. Collected live Internet measurements also show that Westwood+ improves the goodput, from 10% to 46%, with respect to TCP Reno.

An actual trend in the telecommunication world is the proposal of a universal telecommunication system deriving from the convergence between mobile telephony and the Internet. Such a system will presumably exploit both wired and wireless parts, a network platform totally based on a TCP/IP architecture. However, the encapsulation process of TCP/IP layers produces packets whose payload size, for particular services, is a small percentage of the whole packet size (high overhead-services), so that, a significant part of radio channel bandwidth (the most expensive and limited resource of the whole wireless system) is used for header transmission. For these reasons it is of primary importance for the adoption of a header compression scheme to be able to reduce the protocol overhead with the aim to make economically feasible and physically realizable the implementation of high overhead-services. In this paper a new header compression scheme for TCP streams, as a specific header compression profile within the ROHC platform, is proposed and analyzed. This new header compression scheme is based on the distinct management of data and acknowledgment streams associated to the TCP stream. To reduce error propagation, a robust encoding technique (W-LSB) and the synergy between three repair mechanisms for damaged compressed packets have been suggested. Thus, headers up to 40-60 bytes can be compressed to 4-6 bytes with great benefits on cost and quality of service (QoS).

Future cellular mobile communication networks will exploit microcellular architectures and dynamic channel allocation in order to meet the rapidly increasing traffic demand. In this paper, an analytical model has been developed in order to evaluate the performance of maximum packing, a dynamic channel allocation scheme for cellular communication networks. Specifically, a finite number of users has been assumed, moving in a geographical region, covered by a finite set of cells. The considered users are characterized by a variable degree of mobility, which allows both variable sized cells and different user speeds to be analyzed. The model, based on queueing networks, allows the evaluation of the main system performance parameters in terms of blocking probability of new calls, handoff blocking probability, forced termination probability, unsuccessful probability, and throughput. Performance predictions are confirmed by simulation in a wide range of load conditions and user mobility.

Future telecommunication systems will integrate "wireline" and "wireless" networks in a single advanced infrastructure, and it is expected that such systems will exploit asynchronous transfer mode (ATM) technology. This paper considers some aspects of this integrated telecommunication infrastructure. In particular, we propose and analyze an improvement, for wired networks, of the virtual connection tree (VCT) concept, while for the wireless part, a new model for the analysis of a dynamic channel allocation scheme is presented. The obtained analytical results, confirmed by simulation in a wide range of load conditions and user mobility, allow us to evaluate the main system performance parameters, in terms of blocking probability of new calls, handoff blocking probability and forced termination probability.

The main target of this paper is to define and study new multihop architectures capable, potentially of supporting a sustained throughput per user in the order of a Gigabit per second. The Generalized ShuffleNet (GSN) virtual topology is presented and its main characteristics explored, highlighting modularity and rearrangeability. In particular necessary conditions for rearrangeability have been derived analytically, while the sufficiency has been verified, for particular cases, by a computer program. Moreover a novel, table free, routing scheme ensures high performances by distributed hardware based decisions.

We develop a simple and general analytical model for cellular communication networks performance evaluation. The model assumes a finite population of mobiles moving in a finite set of cells. Such a model allows one to evaluate the effects, on system performance, of the following factors: fixed channel allocation (FCA) criteria, load of the users, mobility and distribution of users among cells. The numerical simulation shows that the results of this simple model are acceptable in a wide range of mobility and from light to moderate load conditions. It is pointed out that some of the parameters, that is, the blocking probability of a new call and hand-off blocking probability, are almost invariant with mobility, while forced termination probability and unsuccessful probability do depend on mobility.

In this paper the bandwidth assignment criteria, well suited in ATM LANs environments, have been extensively examined. In particular. After a discussion of the various schemes, we analyzed the bandwidth advertising criterion both by analytical methods and simulation. The cell loss and packet loss probability have been evaluated as a function of the usual network parameters (number of users, packet length, burstiness, etc.), pointing out that the gain factor of the network (i.e., the number of multiplexed sources given the quality of service) is directly proportional to the source burstiness. This intuitive result can be efficiently exploited in ATM LAN environments, where, it is supposed that a significative portion of traffic comes from bursty users.>

The authors propose a redundant, tree structured passive optical network (PON) configuration which can withstand all single and most double link failures. The redundant configuration results from the superposition of two physically separate, diverse-routed trees. This redundancy is transparent to the user. It does not require switching elements or special controls inside the network. It permits easy identification of the fault by the network control center. The fault tolerance properties of the proposed solution are evaluated. Power budget, fiber installation cost, and metropolitan coverage issues are investigated.>

A photonic packet switching communication network is presented and its feasibility is discussed. In particular, it is explicitly supposed that the electro-optical conversion is done only by the end users, while all other network functions (transmission, signal processing, routing, etc.) are implemented by photonic circuits. To interconnect the independent channels obtained by partitioning the global optical bandwidth, the class of topologies, called extended shufflenet, is considered. Two particular networks of this class (multiwavelength loop and extended shufflenet with shared channels), which have routing algorithms simple enough to be implemented by photonic circuits, have been analyzed. The photonic architecture developed for the multiwavelength loop has been applied to the extended shufflenet, which is a feasible possibility for a network with a sustained capacity of the order of the gigabit/s per user. The photonic components used in these circuits have been demonstrated in various laboratories and essentially comprehend delay lines, optical switches, simple optical gates, frequency converters, optical filters, etc. Some of these components have been experimentally integrated in subsystems and the implementation of the proposed circuit with today technology appears feasible.>

The authors consider networks modeled as directed graphs with a uniform link failure probability (q) and develop an algorithm which permits not only a symbolic reliability evaluation, but also a comparison of different topologies to verify if a particular one is uniformly maximally reliable, i.e. it has the best reliability for any value of q. In the case when the previous property cannot be verified, it is possible to evaluate exactly the range of values of q for which one network is better than the other. The method can be extended also to compare networks with different topological characteristics, namely different number of links and/or nodes.>