IEE Colloquium: Twelfth UK Teletraffic Symposium Performance Engineering in Telecommunications Networks
Buy conference proceeding
 Location: Old Windsor, UK
 Conference date: 1517 March 1995
 Conference number: 1995/054
 The following topics were dealt with: modelling techniques and simulation; network modelling; connection acceptance control; ATM buffer management; cell loss modelling; queueing model; network reliability; ATM LAN
1  20 of 23 items found

Some aspects of simulation in telecommunication networks
 Author(s): P. Mars
 + Show details  Hide details

p.
1
(1)
This paper critically reviews certain aspects of the simulation of telecommunication networks. In particular we consider simulation software, the use of highspeed simulation techniques using parallel simulation and the incorporation of artificial intelligence in simulation models. The paper concludes by considering likely future developments in simulation environments including the incorporation of simulation modelling tools as an integral part of the overall network management strategy. (4 pages)

Economies of scale in long and short buffers of large multiplexers
 Author(s): D.D. Botvich ; T.J. Corcoran ; N.G. Duffield ; P. Farrell
 + Show details  Hide details

p.
10
(1)
We examine phenomenological and theoretical aspects of scaling laws of the queuelength distributions in large multiplexers. Anick, Mitra and Sondhi (1982) have shown that the queue length Q^{L} in a multiplexer of L sources served at constant load (i.e. independent of L) has asymptotics for large L of the form p[Q^{L}>Lb]≈e^{LI(b)}, where I is some function (we call it the shape function) depending on the source processes and the load. If I is known, then we can predict the queue length distribution at large L. With a model for the sources, I can be calculated. The procedure for this is outlined. Alternatively we can estimate I. One way to do this is by using the above equation as a definition of I based upon measurements at some L which is both small enough for measurements to be made, but also large enough for the approximation above to be accurate. We illustrate the application of this idea with some empirical queue length distributions obtained by simulation at various L. We find that the most accurate predictions are obtained by taking finitesize effects into account in a simple manner in the approximation. (8 pages)

Dimensioning of shared buffer ATM switch elements
 Author(s): J.A. Schormans ; B.R. Clements ; R.O. Smith
 + Show details  Hide details

p.
11
(1)
ATM communication systems have introduced a number of new dimensioning problems, one being the evaluation of the loss probability for a shared buffer memory within a switch matrix. This is a problem of growing interest since shared buffer memory switches offer the best cell loss probability versus total buffer space provided, and thus are often the design of choice. ATM switches operate in a time slotted fashion: at the end of each slot, N cells will have been transmitted if all N servers have cells for transmission. If a server has no cell to transmit, it will produce an empty slot. Iliadis (1991) has presented a method for solving such a system based on the introduction of an N dimensional Markov chain. Schormans et al. (see Electronics Letters, vol.30, no.1, 1994) have presented a simplified algorithm for a fairly general arrival process. This paper considers certain problems in (shared) buffer dimensioning: calculation of the cell loss probability (CLP) for a 2×2 switching element, given any i/p loading; the effect, on such a shared buffer switching element, of allowing the applied traffic to approach the limit, i.e. during each time slot every input line will contain a new cell; and the relationship between the cell loss probability and the model of the routing strategy. (9 pages)

Buffer management in connectionless servers for AAL5 traffic
 Author(s): A. Asgari ; R. Ahmad ; F. Halsall
 + Show details  Hide details

p.
13
(1)
Connectionless service over ATM networks is supported either, indirectly, by establishing separate end to end virtual connections between ATM end stations or directly by using a connectionless server (CLS). This paper presents an indepth performance comparison of various buffer management strategies in the CLS for AAL5 traffic. All three logical buffers (cellcollection, processing, and departure) in a CLS are considered. Different buffer space allocation schemes and cell discarding mechanisms are studied with respect to their efficiency in buffer space utilisation, frame loss probability and forwarding delays for a given buffer size and traffic. (10 pages)

New for old: reusing ATM cell scale analysis at the burst scale
 Author(s): J.M. Pitts
 + Show details  Hide details

p.
14
(1)
A new burst level queueing analysis has recently been published which challenges the accuracy of fluidflow analysis for application to ATM. It is similar to the conventional approach in that it is carried out at the burst level, but differs in that the excess rate of cells is handled as a discrete process rather than as a continuous fluid. The method gives a mathematically elegant solution when the arrival process is a single on/off source: the state probabilities are a geometric progression, giving a closed form solution for the cell loss probability. This paper summarises how this new discretised fluidflow analysis is being extended by the reuse of cell scale queueing analysis. With a single onoff source, the discretised fluidflow analysis can be considered at the burst level as a queue with geometrically distributed batch input and geometrically distributed service time. The cell scale model provides arbitrarily distributed batch input, with batches arriving as a geometric process. This can be reinterpreted at the burst level as arbitrarily distributed excessrate batches and geometrically distributed service times. Such a model provides an opportunity to specify more closely the excess rate arrival process and apply the analysis to multiplexing scenarios. It also enables multistate sources to be modelled, where each one of many on states is entered from the single off state. Such sources are suitable for modelling output processesthis has potential then for extending the analysis to the network scale. (8 pages)

On worst case traffic in ATM networks
 Author(s): B. Erimli ; J. Murphy ; J. Murphy
 + Show details  Hide details

p.
15
(1)
ATM networks allow for the input traffic from users to vary both from one call to another and within the call. ATM specifies a method for controlling the traffic flow across the user network interface (UNI). This involves each user negotiating call parameters with the network. Once these parameters have been decided then a contract is made between the user and network. The network must then enforce the contract in order to guarantee performance and quality of service to other users. What is of interest to the network is given a particular set of users and contracts what is the worst traffic that the users could input to the network while still abiding by their contracts. This type of input traffic would be called the worst case traffic as it would produce the lowest performance in the network. The reason why this is of importance is that for simulating network performance we would like to have the worst case traffic inputs. Furthermore, it's important for the traffic controller of the network to know the possible worst case traffic so that it can assign parameters accordingly. The contract parameters have been decided upon by the standards organisations and what is needed now is to decide what type of traffic can pass these tests and produce the lowest network performance. This problem has been studied in the literature. We give some theoretical background to explain some of the results in the literature, and we further look at some examples of types of worst case traffic sources. We show that for the two most common types, the greedy onoff and the three state source that either can be worse, depending on the situation. (12 pages)

Discretetime batch renewal processes with applications to the performance modelling of ATM switch architectures
 Author(s): D.D. Kouvatsos and R. Fretwell
 + Show details  Hide details

p.
16
(1)
The indices of dispersion have been proposed for appropriate characterisation of bursty, correlated ATM traffic. However, the analysis of such systems is complex and is often found to be intractable. In this paper it is suggested that, in describing arrival processes for finite buffer queues with deterministic service, batch renewal processes may accurately approximate processes having unbounded indices of dispersion. Analysis of finite buffer queues with deterministic service and batch renewal arrival processes is described and closed form expressions for queue length and loss probabilities are derived. Numerical examples are included in order to examine the effect of correlation of the arrival process on the performance metrics of a multibuffered ATM switch architecture. It is expected that these results will play the role of effective building blocks for the analysis of general ATM networks carrying bursty, correlated traffic. (12 pages)

An improved channel resource queueing algorithm for trunked mobile radio
 Author(s): S.R. Robson and M.C. Sinclair
 + Show details  Hide details

p.
17
(1)
The pooled network resources utilized in a trunked private mobile radio (PMR) network consist primarily of the radio channels and the lineconnected PCM channels which span the fixed basestations. Some calls will not require any landline resource, only a radio path into and out of the local basestation. A single radio channel would meet the resource demands placed on the basestation for calls of this type. However, to set up a crossnetwork communications path between two mobile parties requires several resources which may not be available all at the same time. It is normal practice to allocate the available resources and queue for the remainder. During the busy hour it is not uncommon for both parties involved in the call to queue for the radio channel resource. In such cases the queueing times of the calling and called parties will generally differ. Thus a radio channel may be allocated to one party before the other and consequently remain idle until a radio channel becomes available for the other party. Other calls meanwhile might be unable to proceed for want of that idle radio channel. Bubenik and Turner (1989), for example, reports an increase in carried load in a packet switch as a result of relaxing the strict FIFO queueing discipline. This paper examines the effect of a queuejumping algorithm which bypasses the FIFO discipline and reduces channel idle time. Provided certain criteria are met, it is shown that if one party is moved to the head of a queue as soon as the other party is allocated a radio channel, there is an overall reduction in channel idle time over the entire network. (10 pages)

Cell loss analysis of a multilinear non stationary queuing system using martingale methods
 Author(s): G. Arsenishvili ; T. Kupatadze ; H.F. Rashvand
 + Show details  Hide details

p.
18
(1)
This theoretical work formulates the cell loss probability of a linear queueing system and addresses a generalised analysis of the non stationary statistical characteristics of flow mechanism which can be directly applied to ATM and multimedia communication systems using the martingale approach. (7 pages)

Cost effective analytic solutions for queueing models of ATM networks
 Author(s): D.D. Kouvatsos ; F.A. Zanasi ; J. Wilkinson
 + Show details  Hide details

p.
19
(1)
In the field of ATM technology analytic solutions of proposed queueing models are often hindered by the generation of large state spaces requiring further approximations and a considerable (or even prohibitive) amount of computation. Thus, there is a need to consider alternative methodologies leading to both credible and cost effective performance modelling and evaluation of ATM networks. In the paper closed form solutions, based on entropy maximisation and classical queueing theory, are presented by focusing on the cell loss probabilities of GGeotype independent buffer and shared buffer queueing models. These results are applied to study the performance of an ATM “leaky bucket” policing mechanism, an ATM switch buffer with pushout space priority and also to model an ATM shared buffer switch architecture. Furthermore, these solutions are used, in conjunction with appropriate flow formulae, as building blocks in the decomposition of shared buffer ATM networks into individual ATM switches. Numerical examples involving the leaky bucket mechanism, an ATM switch buffer with space priorities and shared buffer ATM switches are included. (12 pages)

Exploiting concurrency through knowledge of event propagation in cell rate ATM simulation
 Author(s): M. Bocci ; J.M. Pitts ; L.G. Cuthbert
 + Show details  Hide details

p.
2
(1)
A time stepping approach has been adopted by RACE project R2059 ICM for time synchronisation in a parallel event driven ATM network simulator whose purpose is to aid the development of a TMN testbed. This allows the simulation time in all the models distributed across a parallel computing architecture to be kept in step whilst minimising interprocessor communication overhead. Careful setting of the timestep size allows burst length quantisation errors to be minimised. Results are presented which demonstrate that timestepping can allow, efficient parallel simulation whilst keeping burst length quantisation errors within the requirements of the ICM project. However, detailed traffic studies require burst length quantisation to be eliminated altogether as well as improved speedup from parallelising the network model. This paper describes a study into the effectiveness of a two level timestep switching scheme, designed to eliminate burst length quantisation effects whilst maximising the speed of the simulation. The results suggest that, in order to exploit more effectively the concurrency in a spatially decomposed cell rate network simulator an understanding of event propagation on a network level is required. Cell rate analysis shows that events propagate in distinctly differing ways along connections in which the burst scale queueing of cells occurs and along connections in which there is no burst scale queueing. The paper concludes by showing how such knowledge can be used to identify situations in which independent events occur that can be processed concurrently, hence allowing alternative synchronisation schemes. (9 pages)

Design of large switched networks
 + Show details  Hide details

p.
20
(1)
(8 pages)

A distributed protocol for fast and robust virtual path restoration
 Author(s): P.A. Veitch ; D.G. Smith ; I. Hawker
 + Show details  Hide details

p.
21
(1)
Broadband networks must be designed for survivability, otherwise serious failures threaten their integrity. That is why restoration strategies have previously been studied in some depth, first for SDH/SONET transport networks, and then for ATM trunk networks based on the virtual path (VP) concept. In terms of survivability, the characteristics of VPs can be exploited to permit cost effective preassigned restoration by allocating protection paths to working VP routes. This involves setting virtual path identifiers (VPIs) in translation tables of VP crossconnects of the protection route, as well as management of spare capacity reservations in network links. The purpose of this latter operation is twofold: first, spare capacity should be shared between protection paths activated due to different possible failures; second, any failure must not result in the aggregate protection path capacity of any link exceeding the reserved value. In devising a mechanism for activating restoration, the principal goal is to restore failed paths as quickly as possible. This paper describes a novel distributed protocol which enables rapid switching from working to protection routes in the event of serious physical layer failures. A further objective is to instill robustness into the restoration scheme such that adverse effects of less common failures (e.g. node or multiple span) do not degrade network performance. (10 pages)

Analysis of a faulttolerant ATM switch based on a parallel architecture
 Author(s): S. Segkhoonthod and M.C. Sinclair
 + Show details  Hide details

p.
22
(1)
Presents a faulttolerant ATM switch which adapts the idea of using MINs in parallel. It has a distribution network to distribute incoming packets to several routing networks arranged in parallel. Consequently, the input port controller is simply required to submit packets to the distribution network. As the latter is designed to perform some of the routing itself, the routing networks are not required to carry out the whole routing process; consequently incomplete (i.e. truncated) routing networks are used. As a result, the overall complexity of the switch is not further increased by the introduction of the distribution network. The distribution network is also in charge of rerouting packets in the presence of a fault. Unlike other faulttolerant networks where the routing algorithm is usually modified in order to avoid faults, in the proposed switch there is no modification to the routing algorithm in fault conditions: hence the rerouting process is simple and the complexity of switching elements is kept low. Further, it is shown that the proposed switch gives both improved performance and lower complexity than replicated MINs. (9 pages)

Bypassing modelling: an investigation of entropy as a traffic descriptor in the Fairisle ATM network
 Author(s): S. Crosby ; I. Leslie ; J.T. Lewis ; N. O'Connell ; R. Russell ; F. Toomey
 + Show details  Hide details

p.
23
(1)
The thermodynamic entropy of an ATM traffic stream is a useful tool for predicting cellloss ratios and celldelay. It was proposed that the entropy of a cell stream be estimated directly at a multiplexer; from this data, predictions can be made rapidly using algorithms which are simple enough to be executed onthefly. Preliminary investigations using simulated data have suggested that the method is practicable. We describe the results of an investigation of the method using real ATM traffic on the Fairisle ATM network. Fairisle is an experimental ATM LAN whose hardware design permits a high degree of experimental flexibility, including high resolution clocks for timestamping and measurement of traffic, a general purpose CPU and memory on every switch port. Fairisle is equipped with a wide range of ATM traffic sources, including compressed video, audio and LAN data. In addition traffic can be generated from a prerecorded trace of the traffic activity. One such source is the Star Wars movie, reconstructed from a trace of the output of a DCTbased VBR video codec at Bellcore. We describe experiments involving the estimation of the entropy of both singlesource and multiplexed cell streams based on the measured cell interarrival times. These estimates are used to calculate the decayrate of the queuelength distribution. For multiplexed cellstreams, the decayrate can be measured rapidly. (10 pages)

Congestion control in an ATM LAN
 Author(s): F. Ball and D. Hutchison
 + Show details  Hide details

p.
24
(1)
Although originally developed as a standard intended for wide area broadband networks, ATM is now being used in the local environment with ATM LANs appearing well in advance of long haul ATM services. In common with the wide area case, traffic control will play an important role in meeting quality of service (QoS) requirements in an ATM LAN. However, the problems of traffic management in an ATM LAN may be different from those in a wide area ATM network. This paper presents work in progress towards developing a practical solution for call admission control (CAC) in an ATM LAN. A review of congestion control in ATM with emphasis on proposed CAC methods is presented. This is followed by an analysis of the cell level traffic profile expected to be generated by the client workstations. The relevance of existing CAC methods to the local area case is discussed and our proposed initial solution is given. Finally some initial results of simulation studies to verify our solution are presented. (8 pages)

Study of a space priority mechanism for ATM networks using both analytical and simulation techniques
 Author(s): P. Fonseca ; L.G. Cuthbert ; J.M. Pitts
 + Show details  Hide details

p.
3
(1)
Priority mechanisms have been shown to improve the throughput of asynchronous transfer mode (ATM) networks. This paper describes a study of an approximated version of the priority mechanism partial buffer sharing (PBS), using an exact fluid flow analysis and a cellrate ATM network simulator. PBS, a two priority level scheme, discards low priority cells accessing a buffer when it has reached a threshold. The approximated version of PBS does not consider queueing above the threshold, and is applied in both the analytical technique and the cellrate simulator. The cellrate simulation method is used, as it accurately models the burst scale queueing behaviour and leads to speed increases of up to four orders of magnitude, when compared with cell level simulation. Low cell loss probabilities can thus be measured within reasonable computing times. The fluid flow model has the particular characteristic of considering changes in the buffer in discrete steps and it models an onoff traffic source accessing a finite queue. Firstly, the paper explains how the approximated version of PBS was added, both to the analysis and the simulator. It then presents cell loss results obtained with the two approaches, that show the precision of the prioritised simulator. The agreement between the two approaches is particularly significant for low priority cell loss. This study produces a simple and relatively accurate way of helping to foresee the behaviour of a real ATM network in the presence of a priority mechanism. (8 pages)

Routing multiservice traffic in ATM networks
 Author(s): T.P. Jordan ; M.J. Morse ; A. Platt
 + Show details  Hide details

p.
4
(1)
A routing algorithm is proposed for multiservice ATM networks. It is based on a previously proposed algorithm to which significant modifications have been made. The routing algorithm is currently being verified by simulation, but some preliminary comments are made in this paper. Effective bandwidth calculations are assumed to be performed by the network and the routing algorithm must balance the load against a background of long and shortterm calllevel traffic fluctuations. The previously proposed algorithm, minimum loaded path (MLP), computes routes for new calls based on average bitrates of calls and global knowledge of the network status. The adapted algorithm uses effective bandwidths as the basis for routing decisions, and uses flooding to provide each node with knowledge of the network status. The algorithm requires that updates are flooded frequently for effective performance, particularly if the traffic is dominated by shortduration calls. The algorithm performs well at moderate loads but poorly at heavy loads due to the unrestricted use of long, alternative paths. Several different techniques have been proposed to avoid this problem in other network environments, and one or more of these could be adapted for use with the new algorithm. (7 pages)

Complexity classification for performance characterization of system functions within ATM subnetworks
 Author(s): D. Nyong
 + Show details  Hide details

p.
5
(1)
A definition of a generic model for representing communication system behaviour is presented. Also presented is a general complexity measure for classifying computations represented by such a model. The method is based on modelling communication system functions as wreath products of automata which in turn are represented as transformation groups. Emphasis is placed on developing a methodology that can cope with practical ATM functional architectures being proposed by the International Telecommunications Union. Both recursive function complexity and semigroup decomposition complexity theory are interpreted within a common modelling structure. The rule base for the modelling structure is left open, allowing algorithm designers the flexibility of exploiting relational structures from classical algebra. (12 pages)

Resource dimensioning aspect of heterogeneous traffic with different service requirements: integration versus segregation
 Author(s): R. FabregatGesa ; J.L. MarzoLazaro ; P. RidaoRodriguez
 + Show details  Hide details

p.
6
(1)
In this paper, we consider the ATM networks in which the virtual path concept is implemented. The question of how to multiplex two or more diverse traffic classes while providing different quality of service requirements is a very complicated open problem. Two distinct options are available: integration and segregation. In an integration approach all the traffic from different connections are multiplexed onto one VP. This implies that the most restrictive QOS requirements must be applied to all services. Therefore, link utilization will be decreased because unnecessarily stringent QOS is provided to all connections. With the segregation approach the problem can be much simplified if different types of traffic are separated by assigning a VP with dedicated resources (buffers and links). Therefore, resources may not be efficiently utilized because no sharing of bandwidth can take place across the VP. The probability that the bandwidth required by the accepted connections exceeds the capacity of the link is evaluated with the probability of congestion (PC). Since the PC can be expressed as the CLP, we shall simply carry out bandwidth allocation using the PC. We first focus on the influence of some parameters (CLP, bit rate and burstiness) on the capacity required by a VP supporting a single traffic class using the new convolution approach. Numerical results are presented both to compare the required capacity and to observe which conditions under each approach are preferred. (10 pages)