Modern telecommunications networks are being designed to accomodate a heterogenous mix of traffic classes ranging from traditional telephone calls to video and data services. Thus, traffic models are of crucial importance to the engineering and performance analysis of telecommunications system, notably congestion and overload controls and capacity estimation. This chapter surveys teletraffic models, addressing both theoretical and computational aspects. It first surveys the main classes of teletraffic models commonly used in teletraffic modeling, and then proceeds to survey traffic methods for computing statistics relevant to the engineering a teletraffic network. 1 INTRODUCTION Traffic is the driving force of telecommunications systems, representing customers making phone calls, transferring data files and other electronic information, or more recently, transmitting compressed video frames to a display device. The most common modeling context is queueing; traffic is offered to a qu...

High-speed packet networks will begin to support services that need Quality-of-Service (QoS) guarantees. Guaranteeing QoS typically translates to reserving resources for the duration of a call. We propose a statedependent routing scheme that builds on any base stateindependent routing scheme, by routing flows which are blocked on their primary paths (as selected by the state-independent scheme) onto alternate paths in a manner that is guaranteed---under certain Poisson assumptions---to improve on the performance of the base state-independent scheme. Our scheme only requires each node to have state information of those links that are incident on it. Such a scheme is of value when either the base state-independent scheme is already in place and a complete overhaul of the routing algorithm is undesirable, or when the state (reserved flows) of a link changes fast enough that the timely update of state information is infeasible to all possible call-originators. The performance improvements ...

Virtual Path (VP) concept has been gaining attention in terms of effective deployment of ATM (Asynchronous Transfer Mode) networks in recent years. In a recent paper, we have outlined a framework and models for network design and management of dynamically reconfigurable ATM networks based on the virtual path concept from a network planning and management perspective. Our approach has been based on statistical multiplexing of traffic within a traffic class by using a virtual path for the class and deterministic multiplexing of different virtual paths, and on providing dynamic bandwidth and reconfigurability through virtual path concept depending on traffic load during the course of the day. In this paper, we discuss in detail a multi-hour, multi-traffic class network (capacity) design model for providing specified quality-of-service in such dynamically reconfigurable networks; this is done based on the observation that statistical multiplexing of virtual circuits for a traffic class in ...

In this paper, we address a multi-service, multi-rate loss network environment with dynamic routing. In this setting, we consider multiple traffic load periods (multi-hour) during the day, and by observing network dynamics, we present a network dimensioning model that consists of two-steps: a bandwidth estimation estimation step followed by a multi-commodity flow model for multiple services and traffic loads. For network operations, we discuss a probabilistic admission control policy and three multi-service routing schemes. We have used a 10-node network with multiple asymmetric traffic data sets (partially extracted from an actual network) for our study. It was found that the capacity obtained using the analytic network dimensioning model provides a good estimate of network capacity required for meeting the grade-of-service goal for each service type in each traffic load period; this observation is based on a simulated network environment that uses the proposed admission control and t...

We consider multi-class blocking systems in which jobs require a single processing step. There are groups of servers that can each serve a different subset of all job classes. The assignment of jobs occurs according to some fixed overflow policy. We are interested in the blocking probabilities of each class. This model can be used for call centers, tele-communication and computer networks. An approximation method is presented that takes the burstiness of the overflow processes into account. This is achieved by assuming hyperexponential distributions of the inter-overflow times. The approximations are validated with simulation and we make a comparison to existing approximation methods. The overall blocking probability turns out to be approximated with high accuracy by several methods. However, the individual blocking probabilities per class are significantly more accurate for the method that is introduced in this paper.

The burstiness of a traffic stream can have a profound effect on performance statistics in the queue to which the stream is offered, such as line length, waiting times and blocking probabilities. Mathematical descriptors quantifying burstiness should be versatile enough to characterize burstiness, yet analytically tractable to permit numerical calculation. This paper surveys two important descriptors of burstiness: indices of dispersion and peakedness. It discusses their effect on queueing statistics and illustrates them via an example from the domain of TES traffic streams. A conclusion of this example is that peakedness is more general than indices of dispersion. 1 Introduction The effect of burstiness on queueing systems has been long recognized by telephone traffic engineers and other workers [4, 14, 15]. Intuitively, burstiness is present in a discrete traffic process, if its arrival points appear to form visual clusters. The mathematical underpinning of burstiness are less obvi...

Virtual Path (VP) concept is gaining recently attention in terms of effective deployment of ATM networks. In this paper, we present models and algorithms for network design and management of ATM networks based on virtual path concept from a network planning perspective. Our approach is based on statistical multiplexing of traffic within a traffic class by using a virtual path for the class and deterministic multiplexing of different virtual paths, and on providing dynamic bandwidth and reconfigurability through virtual path concept depending on traffic load during the course of the day. For some realistic example networks, we observe that considering network dynamism through variation of traffic during the course of a day by doing dynamic bandwidth and virtual path reconfiguration can save between 10 to 14 % in network design costs compared to a static network based on maximum busy hour traffic. Keywords: Wide-Area ATM Networks; Network planning, servicing and monitoring; optimization...

The exponential growth of Internet brings to focus the need to control such large networks. It is desirable to find simple local rules and strategies that can produce coherent and purposeful global behavior. The need for such mechanisms arises in a variety of settings. In quality-of-service based routing, paths for flows are selected basedupon the knowledge of resource availability at network nodes and the QoS requirements of flows. The best-path routing schemes gather global network state information and always select the best path for an incoming flow based on this global view. On the other hand, the proportional routing schemes proportion incoming flows among multiple candidate paths. We show that near-optimal proportions can be obtained using only locally collected information and a few good candidate paths can be selected using infrequently exchanged global information. We demonstrate that proportional routing schemes achieve higher throughput with lower overhead than best-path routing schemes. Video delivery for full quality playback requires a certain amount of network bandwidth and client buffer. But when these resources are limited, a naive video transmission may cause packet drops at the network and frame drops at the client, resulting in wastage of resources. To avoid this, a server may need to preemptively discard frames locally taking advantage of application-specific information. We first formulate the optimal selective frame discard problem and then present several efficient heuristic algorithms. We also develop adaptive selective layer discard algorithms for providing smoother quality playback of a layered video. Distributed dynamic channel assignment algorithms run at each base station in a wireless cellular network attempt to reduce the network-wide call blocking and call dropping probabilities while making assignment decisions based on neighborhood information only. They may also reassign channels being used by calls in progress to make room for another call. We propose two channel selection strategies based on localpacking for compact packing of channels. We also present a reassignment based call admission control scheme that dynamically adjusts the number of guard channels reserved for handoffs based on reassignment frequency in the neighborhood.

In this paper, we consider renewal arrivals to a communication link with s channels under both a loss system and a finite waiting room system. We assume that such a request arriving to this system holds a channel for exponentially distributed service time. For example, overflow traffic in a teletraffic network, or traffic that is bursty in data networks are non-Poissonian. We characterize renewal arrival by a matrix-exponential distribution (one with a rational Laplace transform). Performance measures of interest for the loss system are the call and time blocking, and for the finite waiting room system the call/time blocking and the average waiting time. Using linear algebraic techniques, we present steady-state balance equations and then describe efficient methods to exactly compute both call and time blocking for both the loss system and the finite-waiting room system. For the loss system, we present numerical results for various distributions considered for the arrival process and compare these with approximations such as Hayward's approximation. For the finite waiting room system, we present results on different sizes of the waiting room and provide comparison with Whitt's approximation. We also discuss results for renewal arrival to a link with buffering in a statistical multiplexing mode. Furthermore, we present results to show that depending on the value of s and the load offered to the system, the third moment of the arrival process may need to be considered for various performance measures.

PhoneFactor is a token-less two-factor authentication service for user remote logons [13]. This allows users of an organization to be authenticated through an automated phone call to the user’s phone before access is allowed. In this paper, we present the service management architecture of PhoneFactor that depends on both the Internet and the public switched telephone network (PSTN), and we identify two key quality of service parameters, the system response time and call blocking probability, where the latter can impact the former. Furthermore, through traffic analysis of the measurement data from the deployed PhoneFactor service, we found that the inter-arrival time of requests follows the Generalized Pareto distribution while the system response time and the call duration (for the authentication part through the phone call) follow the lognormal distribution. Given these distributions, we then present system capacity design methodologies by comparing them to known results for systems that are analytically derivable.