This paper first examines several application areas involving transmission of audio/video content and some pertinent attributes. In the remaining sections, the technical focus is on the prevalent transport of MPEG-2 coded and packetized audio/video content over cell-switched Asynchronous Transfer Mode (ATM) networks. Particular emphasis is placed upon key aspects of interworking between the source coder/decoder devices and the transmission network along with the impact of transport-level impairments upon the resulting quality of service. Although the coverage is specifically oriented toward MPEG-2 over ATM, many of the issues examined should prove relevant to applications involving homogeneous packet-switched (e.g., Internet Protocol) transport or integrated architectures which leverage the respective advantages of cell and packet switching over a multiservice network.

In this paper we propose a framework for the efficient transmission of video traffic through IP-based networks. The IETF's Integrated Services is used to enable the provision of the QoS guarantees required by the video application. Specifically, the Resource ReSerVation Protocol (RSVP) allows the end user to request deterministic QoS guarantees from the network. Since our focus is on pre-recorded video data, the video data is pre-processed through a smoothing operation prior to its transmission. The use of the smoothing algorithm reduces the network resources facilitating the resource reservation process. First, we evaluate the performance of the smoothing algorithm chosen for this study through its sensitivity to processing and network latencies. The second phase of the experimental work consists in evaluating the performance of an RSVP-aware switching point for video transmission supplemented by a smoothing mechanism and a class based queuing scheduler. The overall system evaluation is carried out using various video streams and under different load conditions.

When both TCP and UDP connections co-exist in the Internet environment, the performance of TCP connections is heavily affected by the behavior of `greedy' UDP connections of real-time multimedia applications. In this paper, we propose a new TCP-friendly rate control protocol for video connections, called MPEG-TFRCP, to fairly share the link with TCP connections. To achieve fairness among TCP and UDP connections while performing high quality video transmission, we argue that (1) the interval of rate control must be appropriately determined, (2) the network condition must be accurately predicted, (3) the TCP throughput must be precisely estimated and (4) the video rate must be effectively adjusted. Although our algorithm is based on the existing proposals which do not satisfy all of those conditions, through careful considerations ont he applicability of TFRCP to the actual video applications ours can achieve the high-quality MPEG-2 video transfer while satisfying the TCP-friendliness. Through simulation experiments, we show that the TCP throughput estimation based on pseudo-TCP feedback collection is acceptable and the rate adjustment base don the quantization control should be performed at the interval of 32 times as long as estimated RTT.

Call admission control (CAC) of time-dependent video connections is an important issue for network traffic engineering. The impact of this traffic dependence on video call acceptance region is examined in this article. We considered two different CAC mechanisms; (1) a descriptor- based CAC mechanism and (2) a measurement-based CAC (MBCAC) mechanism. The proposed MBCAC is a hybrid measurement scheme that includes a Kalman filter and a real-time Hurst estimation. We investigated several buffer sizes and video sequences with different dependence degrees. For the accuracy of the Hurst estimation, we developed a Hurst parameter package. The package consists of three different estimators, R/S, Higuchi and Abry-Veitch (wavelet). An important result shows that long-range dependence and short- range dependence connections have similar admission regions.

In this paper, a probabilistic delay model is proposed for the characterization of time-varying nature of the data traffic in the Internet. Connectionist delay representation of the network traffic is derived by considering a videoconference session among a group of N participants connected via various LAN's or randomly through the Internet. Connections among the collaborative participants are represented by their end-to-end network delays involving the physical transmission delay between two stations and the delay due to network traffic. For a network of N stations, this model becomes an N² X N² correlation matrix. When normalized by the delay means and variances of the individual inks, it leads to the N² X N² square matrix of delay cross-correlation coefficients with values varying in the range of [-1,1] independent of scale changes in the correlation amplitude. In the experimental study, three LAN's are selected as the test-bed to measure the random traffic in the Internet. Connections among the different workstations of these three physically separate LAN's are established through the Internet for a long period of observation time. Ensemble averaging over the measurement period dictates that the delay correlation matrix tends to be constant for the selected networks.

In this paper we are considering a mixed platform wireless environment. It means that the mobile spent a Cell Residence Time (CRT) in the originating cell which is a random variable with different statistics from the CRT's in the remaining cells. We model this situation by using a delayed renewal process and we present closed-form mathematical expressions for the probability mass function of the number of handovers during the random interval of a Call Holding Time (CHT). We have modeled the CHT as a mixture of exponential distributions and any distribution for the CRT's, and we establish an interesting relationship between them. In particular, we have considered generalized gamma and circularly distributed CRT's. We also present results for the probability of no completing a call and estimates for the handover rates.

The usage of electrical power distribution networks for voice and data transmission, called Powerline Communications, becomes nowadays more and more attractive, particularly in the telecommunication access area. The most important reasons for that are the deregulation of the telecommunication market and a fact that the access networks are still property of former monopolistic companies. In this work, first we analyze a PLC network and system structure as well as a disturbance scenario in powerline networks. After that, we define a logical structure of the powerline MAC layer and propose the reservation MAC protocols for the usage in the PLC network which provides collision free data transmission. This makes possible better network utilization and realization of QoS guarantees which can make PLC networks competitive to other access technologies.

Reconfiguring a network to counter variations in traffic is expected to greatly enhance optimal usage of network resources. But an important input to this method is the traffic fluctuations themselves. We have developed two models for this purpose to describe the time-dependent variations in traffic at a base station in a nomadic computing, wireless environment. The first model is rather simple and does not take into account details of human behavior. It takes into account the probabilities of choosing different applications. The model is also analyzed and experimented with to identify the important input parameters. The second model, a refined version of the first model, takes into account details of relevant human behavior (in the context of a wireless nomadic computing environment). Finally, we have compared the two models on the basis of their complexity and validity in different situations.

In this contribution, we investigate a discrete-time single- server queue subjected to server interruptions. Server interruptions are modeled as an on/off process with geometrically distributed on-periods and generally distributed off-periods. As message lengths can exceed one time-slot, different operation modes are considered depending on whether service of an interrupted message continues, partially restarts or completely restarts after an interruption. For all alternatives, we establish expressions for the steady-state probability generating functions of the buffer contents at message departure time and at random slot boundaries. From these results, closed- form expressions for various performance measures, such as mean and variance of the buffer occupancy, can be established. As an application, we show that this model is able to assess performance of low-priority traffic in a two- priority HOL scheduling discipline. We then illustrate our approach with some numerical examples.

IP networks are traditionally designed to support a best- effort service, with no guarantees on the reliable and timely delivery of packets. With the migration of real-time applications such as voice onto IP-based platforms, the existing IP network capabilities become inadequate to provide the quality-of-service (QoS) levels that the end- users are accustomed to. While new protocols such as DiffServ and MPLS allow some amount of traffic prioritization, guaranteed QoS requires call admission control. This paper reviews several possible implementations and shows simulation results for one promising method that makes efficient use of the network and is scalable to large networks.

The optimal control of various performance-based measures in high-volume commercial web sites requires a fundamental understanding of the interactions between the diverse set of Internet services that support customer needs and the different importance levels of these services to both the customer and the e-commerce merchant. We present here a study of the control policy for each server in a multiclass queueing network that maximizes a particular function of profit, or minimizes a particular function of cost, across the different classes of Internet services.

Internet2 is a consortium of over 170 universities who are working with industry partners and government agencies to create a high-sped next generation Internet where new research and educational applications and techniques can be tried. The intent is to recreate the spirit of the original Internet to foster innovations in Internet technology and use. Quality of Service (QoS) was built into Internet2 from the beginning. A QoS working group was formed early in the planning stages of Internet2 to define and oversee its QoS approach. This paper gives an overview of the QoS architecture that the QoS working group devised for Internet2, and the resulting Internet2 testbed, known as the Qbone, which the group oversees. The architecture is based on the IETF differentiated services (DiffServ) technique, which was found to meet the requirements that the group formulated. IN particular, the ability of DiffServ to guarantee service levels for individual applications while remaining highly scalable and interoperable was important. Potential uses of the Qbone and some current plans for the Qbone are also discussed.

The demand for QoS provisioning support over Internet grows continuously. The most scalable and less demanding solution, in terms of necessary modifications to the existing Internet infrastructure, is the Differentiated Services (DiffServ) architecture. In this approach, little care has been taken for on a per-application or a per-user basis QoS provisioning. Nevertheless, the Service Level Agreements (SLAs) need to be selected in a way that is efficient both for the users and for the network. In this paper, we develop and evaluate an approach for efficient SLA selection and employing in a DiffServ-over-MPLS network domain. A negotiation process between a user and a network provider is introduced; thus the user can choose from alternative options for allocation of resources the one that better matches his needs. We adopt a usage-based charging scheme that provides the user with the right incentives for SLA selection.

In this paper, we investigated the design and implementation of virtual network service (VNS). We set a virtual private network model on top of which we could create, manage, provide, and evaluated VNS. By using some efficient schemes ensuring QoS requirements, we have built a routing algorithm that service providers would greatly benefit from. In particular, we have used a technology called Blocking Islands to identify the routing paths between sources and destinations. This method has been favorably compared with well-known techniques such as shortest path algorithm and widest path algorithms. Our results will demonstrate that our method is much more efficient in terms of resource allocation for VNS.

We present our research activities and initial results, regarding both simulation and actual test-bed implementation, that aim to demonstrate the feasibility and to evaluate the advantages and performance of a market-based management scheme based on a simple feedback mechanism that informs users of the congestion their traffic is experiencing. The feedback is based on Explicit Congestion Notification marking. The aforementioned objectives are investigated using two service provisioning scenarios. The first scenario involves an application provider that offers discrete Quality of Service (QoS) classes at different prices, whereas the second scenarios offers a wider range of QoS, price pairs. The experiments we report refer to the implementation of a modified version of the TCP algorithm, which provides service differentiation based on a sender's willing-to-pay, and of a packet marking scheme that provides early warnings of incipient congestion.

This paper examines the multi-protocol lambda switching approach to support IP over WDM networks. The MPL(ambda)S approach extends a common control plane to IP and optical domains based on multiprotocol label switching (MPLS) with extensions for the unique characteristics of the optical network. We investigate some practical issues related to signaling, routing, differentiated services, and survivability.

Drawing from the experience and work done in ITU-T on traffic engineering, the feasibility of extending telecommunications network dimensioning methods to IP-based networks with MPLS is explored. Further work is required to develop algorithms for the selection of label-switched paths and for the distribution of traffic to the selected paths.

Considerable interest has arisen in congestion control through traffic engineering from the knowledge that although sensible provisioning of the network infrastructure is needed, together with sufficient underlying capacity, these are not sufficient to deliver the Quality of Service required for new applications. This is due to dynamic variations in load. In operational Internet Protocol (IP) networks, it has been difficult to incorporate effective traffic engineering due to the limited capabilities of the IP technology. In principle, Multiprotocol Label Switching (MPLS), which is a connection-oriented label swapping technology, offers new possibilities in addressing the limitations by allowing the operator to use sophisticated traffic control mechanisms. This paper presents a novel scheme to dynamically manage traffic flows through the network by re-balancing streams during periods of congestion. It proposes management-based algorithms that will allow label switched routers within the network to utilize mechanisms within MPLS to indicate when flows are starting to experience frame/packet loss and then to react accordingly. Based upon knowledge of the customer's Service Level Agreement, together with instantaneous flow information, the label edge routers can then instigate changes to the LSP route and circumvent congestion that would hitherto violate the customer contacts.

To quantify the operational characteristics of a network against a contracted QoS SLA, it is necessary to monitor individual network elements as well as the end-to-end network-level and service-level behavior. Such monitoring technology should also be capable of emulating certain key characteristics of applications that use the network, with the intention of better approximating the user experience. This layered monitoring approach drill down from a service view, to a network view and finally to an individual component element view. In this new layered management model, correlation of fault and performance events becomes the central focus of the solution, along with a meaningful presentation of this correlated information in the context of the services supported across the network.

This paper proposes an approach for the dynamic management of MPLS-based VPNs, MPLS and VPNs significantly contribute to achieve QoS within networks but there remain dynamic management problems associated with their use. We believe that these problems can be solved by using a policy model; such an approach also enables subscribers to keep control of their VPNs and share information with service providers. We used a PCIM-enable network model to account for the peculiarities of the two technologies and combined the resulting schema with COPS and the necessary policy tools. The resulting framework was then tested on a MPLS network. The results show that, with some limitations, the approach does provide the expected functionality.

This paper presents QoStat a GUI based tool, which gives us the possibility to visualize in real time the most important values related to QoS provision. With it we can also change on the fly the most important operational parameters of the mechanisms associated with QoS provision. In this way, it is possible to understand with depth the behavior of those systems--for instance, to study cause-effect relations between the values of the referred parameters and the QoS in fact provided to the different IP traffic flows or classes.

This paper provides comparison of different restoration mechanisms in ATM networks. Each protection mechanism offers different resolution time, and requires different amount of network resources. Our goal was to compare these results and determine which protection mechanism is the best for certain class of user traffic. Comparison is based on results calculated with software tool for designing ATM networks and software tool for simulation of dynamic behavior of ATM networks.

We develop a mathematical model within a game theoretical framework for variable rate real time traffic at a bottleneck node. We address not only the flow control problem, but also pricing and allocation of a single resource among users. A distributed, end-to-end flow control is proposed by introducing a cost function, defined as the difference of pricing and utility functions. For two different utility functions, there exists a unique Nash equilibrium in the underlying game. The paper also introduces three distributed update algorithms, parallel, random and gradient update, which are globally stable under reasonable conditions. The convergence properties and robustness of each algorithm are studied through extensive simulations.

In a sense, the IP DiffServ concept `exploits' important technological trends like growing processing and transmission speeds and decreasing costs of bandwidth. This enables QoS provisioning with less complex traffic management mechanisms that used in other broadband multi- service networks like ATM and IP Intserv. Roughly, one can say that the need for traffic management on small time scales (e.g. sophisticated packet scheduling) becomes smaller, while the role of traffic management on larger time scales (e.g. bandwidth provisioning based on network load measurements) becomes more important. In particular, we argue that admission control in IP DiffServ can be performed on aggregate flow level (important for scalability) instead of on individual flows as in ATM or IP Intserv, while still guaranteeing suitable QoS levels. We also discuss the related issue of network dimensioning and point out possible dimensioning approaches for IP DiffServ networks.

In this paper we present a quality of service routing strategy for network where traffic differentiation follows the class-based paradigm, as in the Differentiated Services framework. This routing strategy is based on a metric of quality of service. This metric represents the impact that delay and losses verified at each router in the network have in application performance. Based on this metric, it is selected a path for each class according to the class sensitivity to delay and losses. The distribution of the metric is triggered by a relative criterion with two thresholds, and the values advertised are the moving average of the last values measured.