An efficient method for computing cell loss probability for heterogeneous bursty traffic in ATM networks

In this paper, we present an approximation for the probability of cell loss of heterogeneous bursty traffic in broadband integrated packet networks based on the asynchronous transfer mode. The sources considered here alternate between active and silent periods and are characterized by their peak and average transmission rates. The cell loss probability is obtained by considering only the number of active sources at a given time and computing the amount of traffic that exceeds the link capacity. Since the amount of buffered excess traffic is not considered in this computation, this approximation is actually an upper bound of the cell loss probability. The numerical efficiency of this bound enables it to be used as a measure based on which robust and simple resource allocation strategies can be developed for bursty sources. Comparison of this bound with the simulation results has shown that the bound is close to the actual loss probabilities especially for large burst lengths and high utilizations.     

Buffer sizing for synchronous self-routeing broadband packet switches with bursty traffic

Previous studies on the performance of synchronous self-routeing packet switches have assumed that the input traffic is random, i.e. there is no correlation between adjacent packet arrivals. This assumption is generally not valid in the data communication environment (e.g. host-to-host communication) where a file transfer usually generates a string of correlated packets. The consequence is that the random traffic assumption greatly underestimates the buffer requirement of the switch. In this paper, we model each input traffic stream as a binary source as a first step to understand the performance of a packet switch in a bursty traffic environment. We found that, given a fixed traffic load (or switch utilization), the required buffer size increases linearly as the burstiness index (the average burst length) of the traffic increases. In addition, the required buffer size is more sensitive to the burstiness of the traffic, when the average traffic load is higher and when the packet loss requirement is more stringent. Initial applications of broadband packet switches are likely to be the interconnections of LANs and hosts. The results of the study indicate that the high burstiness in certain broadband traffic significantly reduces the allowable switch utilization, given a fixed amount of buffers. To increase the switch utilization, an appropriate congestion control mechanism needs to be implemented.     

Comparison of admission control schemes in ATM networks

Asynchronous transfer mode (ATM) offers an efficient means of carrying a wide spectrum of BISDN traffic provided that network congestion is prevented. Unfortunately, efficient congestion control is difficult to achieve in integrated broadband networks, owing to the wide range of traffic characteristics and quality of service (QOS) requirements. We have implemented a network simulator that allows us to evaluate many proposed admission control schemes using many different traffic models. We present the results of several simulation studies, including one study of the performance of the admission control schemes in the presence of traffic sources that exhibit long-term dependence.     

A method for congestion control in atm networks using peak rate throttling

This paper presents a technique, called peak rate throttling, for congestion control of controllable, variable bit rate traffic in ATM (asynchronous transfer mode) networks. It is much less conservative than techniques that have already been proposed. This is because techniques already proposed take no account of the load that is actually being carried by the network. Techniques already proposed assign a predetermined share of network capacity to connections regardless of what is actually being used. The technique proposed here works by throttling the peak bit rate that a user is allowed to send on a connection when the network detects a rising network load that may increase the cell loss rate of connections to greater than an acceptable limit. Users are still permitted to send whatever average bit rate they require, however. The technique relies on the assumption that the load offered to network links is stationary for a long enough period so that controlling action does not need to be performed too often or too quickly. It is feasible to achieve congestion control of a network using peak bit rate throttling alone and no other technique apart from connection control.     

A simple bandwidth management scheme for multiservice ATM networks

In this paper we outline a simple strategy to manage bandwidth allocation and congestion control and to guarantee service requirements in a multiservice ATM (asynchronous transfer mode) network. We propose a network architecture based on a core network, making available a limited number of quality of service profiles, and an access network, made of edge adapters in charge to match the user's requirements with the network's available quality of service profiles and to shape input traffic to achieve high network throughput. The connection admission rule is presented, and edge adapter architecture and design examples are given. The effectiveness of the scheme is demonstrated by numerical results which report the network utilization performance when the proposed core network policies and edge adapter design apply.     

Management of capacity and control in broadband ISDN

The trend towards integration of services onto a single network and the addition of accompanying control features has led to an increase in the complexity of functional signalling protocols. A continuing dramatic reduction in the cost of traffic-carrying capacity has meant that call control costs are becoming an increasingly significant component of the total network cost. In BISDN, the support of many diverse services may result in an excessive control cost. To avoid this potential problem, virtual paths may be used. Capacity may be reserved on virtual paths to simplify the set-up procedure for individual calls. One benefit of this approach is that the cost of establishing a virtual path is shared among all individual connections which subsequently use it. However, this reduced control cost comes at the expense of installing additional traffic carrying capacity. This paper describes techniques for the management of traffic carrying capacity and control mechanisms in BISDN and presents a cost-benefit analysis to determine policies for reserving capacity on virtual paths. Significant savings in the total BISDN cost can result from the use of virtual paths.     

A high-speed-retry banyan switch architecture for giga-bit-rate BISDN networks

This paper proposes a high-speed ATM switch architecture for handling cell rates of several Gb/s in a broadband communication switching system or cross-connect system. The proposed switch architecture, named the high-speed-retry banyan switch, employs a bufferless banyan network between input and output buffers; a cell is repeatedly transmitted from an input buffer until it can be successfully transmitted to the desired output buffer. A simple cell-retransmission algorithm, is employed as is a ring-arbitration algorithm for cell conflict. They are suitable for FIFO type buffers and bufferless highspeed devices. Good traffic characteristics which are independent of switch size are achieved for an internal speed ratio of only four times the input line speed. A prototype system with the internal speed of 1·2 Gb/s is constructed in order to confirm the basic operation of the high-speed-retry banyan switch. The prototype system, even in its present state, could be used to realize a giga-bit-rate BISDN switching system.     

Expanse software for distributed call and connection control

The experimental ATM network services environment (EXPANSE) prototype provides an experimental testbed for multimedia multiparty telecommunications services over heterogeneous communications networks. The EXPANSE software architecture supports the functional separation of call and connection control. At the call control layer an object-oriented, transaction-based call model provides for the co-ordination of the negotiations among users for the establishment, modification and disconnection of telecommunications services. The connection control layer provides an object-oriented protocol for the dynamic establishment, modification and release of network connections which include the control and allocation of network resources required for multimedia multiparty services. The EXPANSE software includes an application programming interface (API) to the signalling protocol at the call control level. The API provides a uniform and powerful interface to the call model and allows different applications to co-ordinate the control of local resources and session state.     

A three-stage architecture for very large packet switches

This paper proposes a three-stage broadband packet switch architecture with more than 16,000 ports for a future central office. The switch is constructed by interconnecting many independent switch modules of small size which can be implemented using modifications of various well-studied switch fabric designs. Channel grouping is used to provide multiple paths for each input-output pair in order to decrease delay and increase throughput. We show that, for a given size, switch modules with channel grouping are simpler to realize than those without channel grouping. A datagram packet routeing approach is adopted in order to avoid table look-up that would be required by virtual-circuit routeing. Ways of preserving the sequence integrity of packets under this situation are presented. Performance analyses show that a 32,768 x 32,768 switch with acceptable performance can be constructed based on switch fabrics of no more than 128 ports.     

A call admission control strategy for multiservice wireless cellular packet networks

A simple connection control system for multiservice cellular wireless networks is presented. Mobile stations are classified depending on the traffic they generate (e.g., voice, data). Within each class, two subclasses are also identified: stations which have originated inside the cell and stations which come from adjacent cells. The connection control mechanism is carried out by considering a number of priorities among the various classes and their subclasses. It works on two levels: static and dynamic. The static level looks at packet-level quality of service (QoS), such as cell loss and delay, while the dynamic level takes care of connection dynamics and allows the load of the system to be driven with respect to the various subclasses. Results that illustrate the performance of this control mechanism are presented.     

Experimentation on asynchronous switching techniques for broadband ISDN

This paper deals with a proposed implementation exploiting the new asynchronous transfer mode (ATM) multiplexing and switching techniques. In particular, the basic architectural and functional choices are outlined and various aspects are examined regarding the implementation of a switching arrangement using ATM switching techniques, giving some indication about its hardware and software architecture and the foreseeable system performance.     

Statistical multiplexing gain for variable bit rate video codecs in ATM networks

Asynchronous transfer mode (ATM) broadband networks will support variable bit rate video codecs, which are capable of maintaining a constant picture quality. To demonstrate this capability, a prototype hardware video coder has been developed in the Siemens Central Communications Laboratories. The prototype uses interframe coding, combined with a discrete cosine transform, and is able to reproduce the original picture quality, independent of signal sources or picture material used. A gain in transmission efficiency is expected when several video sources share a common ATM channel (‘statistical multiplexing’). This paper reports on a series of measurements that have been performed using this coder for a large variety of video sources to determine the possible gain in transmission efficiency. The main results are: for realistic video phone scenes, up to about three times the number of signals can be transmitted compared to transmision with constant rate and the same picture quality, if the output signal of the coder has been smoothed over a period of one frame. Smoothing over shorter periods reduces the potential gain substantially. The statistical multiplexing gain increases with the duration of the picture sequences due to the criterion of constant picture quality. It varies very little with the acceptable packet loss rate.     

OAM mechanisms for BISDN UNI and access networks

This paper clarifies operation and maintenance (OAM) requirements for broadband integrated services digital network (BISDN) customer access systems including user-network interface (UNI), and proposes basic OAM mechanisms. Access networks employ a logical networking architecture based on the asynchronous transfer mode (ATM) virtual path (VP) concept to provide all services efficiently, economically. Services are independent of the network's physical configuration. Thus, OAM functions for the access network are required to support physical and ATM layer capabilities. UNI physical performance and failure information flows to ensure user service quality are described. A fault localization mechanism is introduced that determines whether the failure lies on the network provider side or customer side. It uses failure information flows and loop back testing at the network terminator 1 (NT1). As connection-related virtual path connection (VPC) availability indication is necessary for user-to-network applications and user-to-user applications, two alternative VPC-related availability indication mechanisms are studied. Furthermore, basic performance monitoring mechanisms for ATM layer are described.     

A high-speed four-channel subscriber fibre-optic link for broadband networks

This paper describes the performance of a 560 Mb/s digital fibre-optic link designed specifically for possible use in a broadband integrated services digital network (BISDN) environment. Four 140 Mb/s non-return-to-zero (NRZ) channels are time division multiplexed using a bit-by-bit multiplexing strategy and transmitted using potentially low cost 1.3 μm EE-LED and laser sources. A 10^?9 bit error rate with a power level of -34 dBm was achieved for each independent channel over a distance of 10 km.     

A high-capacity broadband packet switch architecture based on a multilink approach

Broadband packet networks based on asynchronous transfer mode (ATM) are expected to provide a wide range of services, including motion video, voice, data and image. When these networks become prevalent, some applications such as motion video and high-speed LAN interconnections will place a very large bit rate requirement on the channels. Currently, the physical layer supported by the synchronous optical network (SONET) allows the transmission of up to 2.4 Gbit/s with the OC-48 optical interface. However, it is not feasible for the electronic packet switch to route packets at this rate on a single link. In this paper we present a design of a broadband packet switch that uses multiple links in parallel to realize a high-speed channel. This implementation permits the switch to operate at the lower link rate, which can be at 150 Mbit/s, while having the ability to support a virtual circuit at a higher rate (up to 2.4 Gbit/s). The main contribution of the design is that packet sequence on a channel is still maintained even though packets are allowed to use any of the links belonging to the same channel. Besides allowing the switch to function at a slower rate than the transmission channel rate, the implementation of the multilinks benefits from statistical multiplexing gain. Analytical results show the performance advantages of multilink design with respect to delay, throughput and packet loss probability.     

Assembly admission control based on random packet selection at border nodes in Optical Burst-Switched networks

In Optical Burst-Switched (OBS) networks employing Just Enough Time (JET) signalling, Burst-Control Packets request resources at intermediate nodes following a one-way reservation protocol, that is, requested resources are not confirmed back to the source. Hence, data bursts are transmitted without any guarantees, and it sometimes occurs that these are dropped at a certain hop in the source–destination path, hence wasting resources at previous hops. This effect is specially harmful if some connections are abusing of the global shared resources, violating their respective Service Level Agreements, thus causing: (1) global performance degradation; and, (2) unfair service received by other connections. This article proposes “Random Packet Assembly Admission Control”, an admission control mechanism for OBS networks that moderates the two problems above. The mechanism monitors the network load status, detects which links are heavily loaded and decides which flows among the total traversing them require throughput decrease, on attempts to alleviate congestion and benefit other flows which are not abusing from the network. Such throughput decrease consists of preventive packet dropping during the assembly process at the ingress nodes of the OBS network, thus making no use of the network core. The numerical results show a substantial increase in the throughput experienced by well-behaved flows, and fundamental fairness achievement in the use of optical resources.

Virtual spacing for flexible traffic control

Weighted fair queuing emerges as an elegant and effective solution to the problem of integrating real time services with strict delay constraints and high-speed data services for which significant delays cannot always be avoided. Virtual spacing is a weighted fair queuing scheduling algorithm intended to be applied in the context of the ATM based BISDN. After discussing how virtual spacing allows flexible traffic control for a variety of service types, we discuss the issue of its implementation in switching nodes. The aim of the paper is to highlight the advantages of weighted fair queuing in enabling a truly integrated services network and to propose a technologically feasible implementation.     

Performance of in-call buffer/window allocation schemes for short intermittent file transfers over broadband packet networks

While static open loop rate controls may be adequate for handling continuous bit rate (CBR) traffic, relatively smooth data traffic, and relatively low speed bursty data traffic over broadband integrated networks, high speed bursty data sources need more dynamic controls. Burst level resource allocation is one such dynamic control. Potential benefits and other issues for burst level resource parameter negotiations for bursty data traffic over high speed wide area packet networks have been discussed earlier.^1–6 A detailed analysis of an adaptive buffer/window negotiation scheme for long file transfers using these concepts is presented in Reference 1. In this paper we discuss two burst level buffer/window negotiation schemes for short intermittent file transfers, focusing on the specific needs of such traffic streams. We develop closed network of queues models to reflect the behaviour of the proposed schemes. These models, while being simple, capture essential details of the control schemes. Under fairly general assumptions, the resulting network of queues is of product form and can be analysed using the mean value analysis. We use such an analysis to compare the proposed schemes and to determine appropriate sizes of trunk buffers to achieve the desired balance between bandwidth utilization and file transfer delay. The effects of other parameters on the performance of these schemes as well as on the buffer sizing rules are also discussed. Burst level (in-call) parameter negotiation may be carried out by the end system with the network elements or by an interface system (access controller) with the broadband network elements. We discuss implications of this location as well as the needed protocol features. Finally, the service discrimination capabilities desired at the trunk controllers in switching nodes are briefly discussed.     

Traffic studies of a multiplexer in an ATM network and applications to the future broadband ISDN

In the first part of this paper the influence of the arrival process on these three performance parameters is discussed. The values of these three performance parameters depend strongly on the utilization distribution over the input trunks of the multiplexer while the total utilization on the input trunks is fixed. If there is the same utilization on all input trunks, the values of these three performance paramters reach a maximum. This result is also valid for an ATM switching matrix with a separate FIFO buffer at each output trunk. In the second part of this paper some examples are provided of how to apply the results of these studies to the future broadband ISDN which is intended to transmit services with very high bit rates (e.g. 135 Mb/s) as well as services with very low bit rates (e.g. 64 kb/s).