High-speed ATM Networks Research Articles

Scalable ATM network interface design using parallel RISC processors architecture

In this work, a high-speed scalable ATM network interface has been designed and simulated. The interface has two processing engines, one for the transmission and the other for receiving side. Two specialized single-issue RISC cores supported with three stages pipeline and forwarding engine, have been used to process the network interface functions and ATM protocols. In addition, the interface architecture has Content Addressable Memory, five First-In-First-Out buffers, two simple Direct Memory Access units, two dual-port RAM, host interface, and transmission lines interface. The performance evaluation of the network interface has been measured through the development of a VHDL-based cycle accurate simulator. The results have shown that such network interface is scalable and could support a wire-speed of 2.4 Gb/s when the RISC cores run at a clock rate of about 170 MHz.

Multimedia call admission control in mobile networks: a dynamical reservation-pool approach

The effective delivery of Multimedia traffic in next generation mobile cellular communication environment will become an increasingly important issue as cell sizes shrink to accommodate continuously large demand for high capacity. To provide the guaranteed mobile QoS for arriving multi-class calls, we should not only consider the regular latency requirement, but also minimize the dropping rate of handoff calls while at the same time controlling the blocking rate of new calls. In this paper, a new multi-class call admission control mechanism is proposed that is based on dynamically forming a reservation-pool for handoff requests. The prerequisite of this scheme is that wireless components can estimate future trajectory of mobile hosts with high degree of accuracy, which is possible in today’s increasingly improved position location techniques such as GPS. In this work, we first develop the notion of servicing various multi-class connections based on priority determined using the combined analysis of mobile movement information and the desired QoS requirements of multimedia traffic. We also provide a practical framework for determining the occurrence time of handoff request reservations. For our simulation experiments, three kinds of timers are introduced for controlling the progress of discrete events. The simulation results show that the individual QoS criteria of multi-class traffic such as the handoff call dropping probability can be achieved within a targeted objective and the new call blocking probability is constrained to be below a given level. The proposed scheme is applicable to channel allocation of multi-class calls over high-speed multimedia wireless ATM networks.

Performance Analysis of Banyan Switch Architecture

The advancement in the field of computer technology, communication and the increasing user demand in terms of bandwidth and speed, the future communication network is rolling towards the BISDN with optical technology. The high speed ATM network is one of the major tool used for BISDN. In turn, the efficiency of the ATM network relies on the switching fabric. Hence, the switching fabric of future telecommunication systems should have a very low cell loss and delay performance. In this paper, the simulation results of Sort Banyan architecture of size 8 × 8 are presented. The simulation of Sort Banyan network is performed along with four different methods in view of enhancing the throughput of the Sort Banyan switch network. The cell loss and delay performance versus the cell arrival rate for the four schemes viz, Input buffer, Interstage buffer, Shared output buffer and Regeneration of cells using FEC method are presented for both Poisson and Self-similar traffic models.

The Object-Oriented Design and Performance of JAWS: A High-performance Web Server Optimized for

This paper provides two contributions to the study of high-performance object-oriented (OO) Web servers. First, it outlines the design principles and optimizations necessary to develop efficient and scalable Web servers and illustrates how we have applied these principles and optimizations to create JAWS. JAWS is a high-performance Web server that is designed to alleviate overheads incurred by existing Web servers on high-speed networks. In addition to its highly extensible OO design, it is also highly efficient, consistently outperforming existing Web servers, such as Apache, Java Server, PHTTPD, Zeus, and Netscape Enterprise, over 155 Mbps ATM networks on UNIX platforms. Second, this paper describes how we have customized the JAWS OO design to leverage advanced features of Windows NT on multi-processor platforms linked by high-speed ATM networks. The Windows NT features used in JAWS include asynchronous mechanisms for connection establishment and data transfer. Our previous benchmarking studies demonstrate that once the overhead of disk I/O is reduced to a negligible constant factor (e.g., via memory caches), the primary determinants of Web server performance are its concurrency and event dispatching strategies. Our performance results over a 155 Mbps ATM network indicate that certain Windows NT asynchronous I/O mechanisms (i.e., TransmitFile) provide superior performance for large file transfers compared with conventional synchronous multi-threaded servers. Conversely, synchronous event dispatching performed better for files less than 50 Kbytes. Thus, to provide optimal performance, a Web server design should be adaptive, i.e., choosing to use different mechanisms (such as TransmitFile) to handle requests for large files, while using alternative I/O mechanisms (such as synchronous event dispatching) on requests for small files.

The design and performance of MedJava. Experience of developing performance-sensitive distributed applications with Java

The Java programming language has gained substantial popularity in the past two years. Java's networking features, along with the growing number of Web browsers that execute Java applets, facilitate Internet programming. Despite the popularity of Java, however, there are many concerns about its efficiency. In particular, networking and computation performance are key concerns when considering the use of Java to develop performance-sensitive distributed applications. This paper makes three contributions to the study of Java for performance-sensitive distributed applications. First, we describe an architecture using Java and the Web to develop MedJava, which is a distributed electronic medical imaging system with stringent networking and computation requirements. Second, we present benchmarks of MedJava image processing and compare the results with the performance of xv , which is an equivalent image processing application written in C. Finally, we present performance benchmarks using Java as a transport interface to exchange large medical images over high-speed ATM networks. For computationally-intensive algorithms like image filtering, Java code that is optimized both manually and with JIT compilers can sometimes compensate for the lack of compile-time optimizations and yield a performance commensurate with equivalently compiled C code. With rigorous compile-time optimizations, however, C compilers still generally generate more efficient code. The advent of highly optimizing Java compilers should make it feasible to use Java for performance-sensitive distributed applications where C and C++ are currently used.

Guest Editor's introduction: Selected papers from the 4th USENIX Conference on Object-Oriented Technologies and Systems

Hewlett-Packard Laboratories, 1501 Page Mill Road, Palo Alto, CA 94304, USA Introduction The USENIX Conference on Object-Oriented Technologies and Systems (COOTS) is held annually in the late spring. The conference evolved from a set of C++ workshops that were held under the auspices of USENIX, the first of which met in 1989. Given the growing diverse interest in object-oriented technologies, the C++ focus of the workshop eventually became too narrow, with the result that the scope was widened in 1995 to include object-oriented technologies and systems. COOTS is intended to showcase advanced RD the architecture and implementation of distributed object systems (e.g. CORBA, DCOM, RMI); object-oriented programming and specification languages; object-oriented design and analysis. The 4th meeting of COOTS was held 27 - 30 April 1998 at the El Dorado Hotel, Santa Fe, New Mexico, USA. Several tutorials were given. The technical program proper consisted of a single track of six sessions, with three paper presentations per session. A keynote address and a provocative panel session rounded out the technical program. The program committee reviewed 56 papers, selecting the best 18 for presentation in the technical sessions. While we solicit papers across the spectrum of applications of object-oriented technologies, this year there was a predominance of distributed, object-oriented papers. The accepted papers reflected this asymmetry, with 15 papers on distributed objects and 3 papers on object-oriented languages. The papers in this special issue are the six best distributed object papers (in the opinion of the program committee). They represent the diversity of research in this particular area, and should give the reader a good idea of the types of papers presented at COOTS as well as the calibre of the work so presented. The papers The paper by Jain, Widoff and Schmidt explores the suitability of Java for writing performance-sensitive distributed applications. Despite the popularity of Java, there are many concerns about its efficiency; in particular, networking and computation performance are key concerns when considering the use of Java to develop performance-sensitive distributed applications. This paper makes three contributions to the study of Java for these applications: it describes an architecture using Java and the Web to develop MedJava, which is a distributed electronic medical imaging system with stringent networking and computation requirements; it presents benchmarks of MedJava image processing and compares the results to the performance of xv, which is an equivalent image processing application written in C; it presents performance benchmarks using Java as a transport interface to exchange large medical images over high-speed ATM networks. The paper by Little and Shrivastava covers the integration of several important topics: transactions, distributed systems, Java, the Internet and security. The usefulness of this paper lies in the synthesis of an effective solution applying work in different areas of computing to the Java environment. Securing applications constructed from distributed objects is important if these applications are to be used in mission-critical situations. Delegation is one aspect of distributed system security that is necessary for such applications. The paper by Nagaratnam and Lea describes a secure delegation model for Java-based, distributed object environments. The paper by Frolund and Koistinen addresses the topical issue of providing a common way for describing Quality-of-Service (QoS) features in distributed, object-oriented systems. They present a general QoS language, QML, that can be used to capture QoS properties as part of a design. They also show how to extend UML to support QML concepts. The paper by Szymaszek, Uszok and Zielinski discusses the important issue of efficient implementation and usage of fine-grained objects in CORBA-based applications. Fine-grained objects can have serious ramifications on overall application performance and scalability, and the paper suggests that such objects should not be treated as first-class CORBA objects, proposing instead the use of collections and smart proxies for efficient implementation. The paper by Milojicic, LaForge and Chauhan describes a mobile objects and agents infrastructure. Their particular research has focused on communication support across agent migration and extensive resource control. The paper also discusses issues regarding interoperation between agent systems. Acknowledgments The editor wishes to thank all of the authors, reviewers and publishers. Without their excellent work, and the contribution of their valuable time, this special issue would not have been possible.

First-order rate-based flow control with dynamic queue threshold for high-speed wide-area ATM networks

In this paper we present a new rate-based flow control scheme for ATM ABR services and analyze its performance. The proposed algorithm, which we refer to as first-order rate-based flow control (FRFC) is the most simple form of queue-length-based flow control. The asymptotic stability, the steady-state throughput, queue length and fairness, and the transient behavior are analyzed for the case of multiple connections with diverse round-trip delays. We also consider a novel approach to dynamically adjust a queue threshold in the FRFC according to the changes in the available bandwidth, and the arrival and departure of connections. Simulations show that the simple FRFC with dynamic queue threshold (DQT) effectively maintains high throughput, small loss and a desired fairness in these dynamic environments and is a promising solution for ABR flow control in ATM networks.

Modelling And Design Of Dependable High-Speed Information Networks

Unlike conventional packet-switched networks, the ATM network is expected to perform connection-oriented operations for calls with variable traffic characteristics. Fault tolerance in highspeed ATM networks it greatly needed even for short-time link failures due to the possibility of large cell. In this paper, we develop an ATM backbone network design algorithm and network resource management schemes to cope with network fault tolerance by considering different qualities of service, such as cell loss probability and mean end-to-end delay requirements for each call request. In addition, we ensure that these performance requirements are met even under faulty network conditions. Numerical examples show that our algorithm yields good solutions.

Distributed clinical neurophysiology.

We have developed a consultation forum for clinical neurophysiology in Finland. The system connects local digital electroencephalography (EEG) recording and analysing networks using a high-speed asynchronous transfer mode (ATM) network. Clinicians can obtain a second opinion using interactive data and video consultations or using data-only consultations. In addition, the system can be used for off-line review of pre-recorded data. During a one-month evaluation, 66 EEG recordings were made altogether in Satakunta Central Hospital and consultations were required on 12 occasions. Nine of them were data-only consultations and three were data and video consultations. A data consultation lasted 15-20 min and a data and video consultation 35-45 min. Clinically, there were numerous benefits for the hospitals. The system established a link to a centre of excellence for second opinions or continuing education. It also helped with on-duty arrangements and enabled the construction of national data banks.

Performance benchmarking of signaling in ATM networks

The overhead associated with establishing switched virtual circuits (SVCs) in high-speed wide area ATM networks is an important performance factor in applications with client/server architectures, including the World Wide Web and low-latency applications such as command/control and modeling/simulation. This article quantifies this overhead by describing a benchmarking framework implemented in a reliable, portable, and easily expandable toolset designed to conduct measurements of well-defined parameters and performance metrics on equipment that is either isolated or connected to a live network. The usefulness of this toolset is then demonstrated in a test suite applied to assess call rates and connection establishment latency of point-to-point and point-to-multipoint ATM connections in the MAGIC, ATDnet, and AAI WANs, on the campus ATM LAN at NRL, and on individual switches in our laboratories.

Measuring the performance of communication middleware on high-speed networks

Conventional implementations of communication middleware (such as CORBA and traditional RPC toolkits) incur considerable over-head when used for performance-sensitive applications over high-speed networks. As gigabit networks become pervasive, inefficient middleware will force programmers to use lower-level mechanisms to achieve the necessary transfer rates. This is a serious problem for mission/life-critical applications (such as satellite surveillance and medical imaging).This paper compares the performance of several widely used communication middleware mechanisms on a high-speed ATM network. The middleware ranged from lower-level mechanisms (such as socket-based C interfaces and C++ wrappers for sockets) to higher-level mechanisms (such as RPC, hand-optimized RPC and two implementations of CORBA - Orbix and ORBeline). These measurements reveal that the lower-level C and C++ implementations outperform the CORBA implementations significantly (the best CORBA throughput for remote transfer was roughly 75 to 80 percent of the best C/C++ throughput for sending scalar data types and only around 33 percent for sending structs containing binary fields), and the hand-optimized RPC code performs slightly better than the CORBA implementations. Our goal in precisely pinpointing the sources of overhead for communication middleware is to develop scalable and flexible CORBA implementations that can deliver gigabit data rates to applications.

Characterizing highly correlated video traffic in high-speed asynchronous transfer mode networks

The enormous bandwidth potential of optical fiber has resulted in a worldwide effort to develop high-speed ATM networks, also called broadband integrated services digital networks (BISDN). Many of the applications that ATM networks will support will have a strong video component to them. Hence, it is important to understand the behavior of video traffic as it travels through these networks. To that end, we develop the generalized histogram model (GHM) to characterize ''highly correlated'' traffic, such as motion JPEG or ''smoothed'' MPEG traffic over ATM networks end-to-end. Using our GHM model we show how to determine the loss rate at any node in an ATM network. We find that, for highly correlated video sources, increasing the buffer size beyond a certain region called the ''cell region'' only marginally decreases the probability of loss. This implies that large buffers cannot be used to control the loss for such sources. The analytical model provided in this paper can be used for admission control, and network dimensioning and design in ATM networks. We have validated our results using simulations of real traces of video sources.

Improvements to TCP performance in high-speed ATM networks

Performance of TCP connections in high-speed wide-area ATM networks is of great importance due to the widespread use of te TCP/IP protocol for data transfers and the increasing deployment of ATM networks. ATM technology promises substantial gains by allowing the available link bandwidth to be efficiently used by a mix of audio, video, and data traffic. However, certain features of ATM technology lead to a substantial degradation of TCP connection throughput under congestion conditions. The poor performance of TCP has been observed in a number of recent studies [7] and a number of remedies have been proposed [1, 8], but most of the additional mechanisms proposed to improve TCP performance in ATM networks require substantial complication of the ATM switches

A two-server queueing system with periodic and credit-based server availabilities

A discrete-time, two-server queueing system is studied in this paper. The service time of a customer (cell) is fixed and equal to one time unit. Server 1 provides for periodic service of the queue (periodT). Server 2 provides for service only when server 1 is unavailable and provided that the associated service credit is nonzero. The resulting system is shown to model the queueing behavior of a network user which is subject to traffic regulation for congestion avoidance in high speed ATM networks. A general methodology is developed for the study of this queueing system, based on renewal theory. The dimensionality of the developed model is independent ofT;T increases with the network speed. The cell loss probabilities are computed in the case of finite capacity queue.

Guaranteed approximation of Markov chains with applications to multiplexer engineering in ATM networks

Discrete Markov chains are applied widely for analysis and design of high speed ATM networks due to their essentially discrete nature. Unfortunately, their use is precluded for many important problems due to explosion of the state space cardinality. In this paper we propose a new method for approximation of a discrete Markov chain by a chain of considerably smaller dimension which is based on the duality theory of optimization. A novel feature of our approach is that it provides guaranteed upper and lower bounds for the performance indices defined on the steady state distribution of the original system. We apply our method to the problem of engineering multiplexers for ATM networks.

Signaling and operating system support for native-mode ATM applications

Applications communicating over connectionless networks, such as IP, cannot obtain per-connection Quality of Service (QoS) guarantees. In contrast, the connection-oriented nature of the ATM layer and its per-virtual-circuit QoS guarantees are visible to a native-mode ATM application. We describe the design and implementation of operating system and signaling support for native-mode applications, independent of the semantics of the protocol layers or of the signaling protocol. The work was done in the context of a Unix-like operating system and the Xunet 2 wide-area high-speed ATM network. The IPC-based interface between an application and the signaling entity allows processes to request parameterized virtual circuits, and the signaling-kernel interface allows resources to be reclaimed from prematurely terminating processes. We also built a simple encapsulation layer over raw IP that allows any host with IP access to send AAL frames into the wide-area network with little performance degradation. Our design makes it simple to port existing TCP/IP socket applications to a native-mode ATM protocol stack and also enables interoperation of existing IP networks with our ATM network. Our experience has been positive - the design is robust, easily extendible and scales well with the number of open connections.

Analysis of out-of-sequence problem and preventative schemes in parallel switch architecture for high-speed ATM network

The parallel switch architecture combines advantageously a parallel plane method with a bit-parallel conversion scheme in VLSI implementation of ATM switches with the capability of handling unlimited external trunk speed. However, the parallel architecture encounters out-of-sequence problems owing to various possible switching paths. The authors analyse the out-of-sequence performance dependency on the trunk utilisation, the total number of virtual channels/trunk, and the number of parallel switch planes. They then propose two parallel switches equipped with preventive schemes for avoiding cells delivered out-of-sequence. By using the method of cascading virtual channel routing networks at input lines of the switch planes, the large speed reducing effect to support trunks at speeds of several Gb/s can be achieved in ATM-based parallel switch design. Also, the out-of-sequence problem is basically eliminated, and additional delay can be reduced to a negligible level.

A counter based congestion control (CBC) for ATM networks

Abstract This paper presents and analyzes a new congestion control strategy targeted toward integrated services in high speed ATM networks. The proposed control combines a new transmission control scheme with an existing access control to provide efficient, fair, and congestion-free network control. The transmission control scheme uses counters at each node to regulate the flow of packets from the output packet queues to the outgoing link. The transmission control is designed to be flexible in accomodating various existing and expected applications and to be simple in implementation. The resulting congestion control strategy supports different service rate for each service class according to its individual requirements and meets the GoS of each service class. The strategy is proven to provide bounded end-to-end queueing delay for each individual real-time application and at the same time give a best effort service to loss-sensitive and delay-tolerable data streams. An analytical model is presented to study the system state queueing behavior and the results how that the proposed strategy also has a good average performance.

Performance analysis of virtual circuit connections for bursty data sources in ATM networks

ATM is a packet-like transmission mode that has been proposed for BISDN. It is characterized by an asynchronous slotted transmission mechanism that provides a high bandwidth, low delay connection-oriented transport service to the end user. In this paper, we provide an analytical approach for determining the performance of a virtual circuit connection for data transmission in a high-speed ATM network with finite buffers at the network nodes. The analysis assumes that the network operates using thebest effort delivery strategy and that the end-to-end virtual circuit is responsible for guaranteeing the integrity of the connection. Since the normal Markovian assumptions do not apply, a concise exact solution is impossible to obtain. This provides motivation for developing approximate techniques such as those found in Whitt's QNA paper that allow us to use general distributions for the traffic streams and service times. However, even these techniques assume infinite buffer capacities and hence cannot model buffer overflow. We have therefore developed a hybrid model that allows us to incorporate finite buffers at the nodes. This enables us to study the effect on the performance of both link errors and buffer overflow in conjunction with an end-to-end packet loss recovery scheme.