Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. An Investigation into Multimedia Local Area Networks A Thesis presented in partial fulfillment of the requirements for the degree of Master of Technology in Information Engineering at Massey University SUSARLA UDA YA KUMAR 1997. Acknowledgement I express my deep sense of gratitude to my supervisor Dr. Jamil Y Khan, for all his help and guidance throughout this study. I also express my deep gratitude to my co-supervisor Prof. R.M. Hodgson for all his help and encouragement. I also thank Prof. Jagan P Agrawal and Dr. Upkar Varshney for their prompt clarifications on their paper entitled "Performance Evaluation of a Multimedia Local ATM Network". I also thank my family members for their unstinted support and encouragement during this study. Finally, I thank all those who are directly or indirectly helped me in this study. 11 Abstract In this thesis the performance of the Multimedia Local Asynchronous Transfer Mode Network (MLAN) protocol is evaluated by a computer simulation method using voice and data source models. SIMSCRIPT II.5, a discrete event simulation language is used for the simulation. In addition, Fiber Distributed Data Interface (FDDI) and Fast Ethernet networks were simulated for data traffic and their performance is evaluated using COMNET III, a communication network simulation package. The main aim of this work is to evaluate the performance of the MLAN and to analyse the suitability of MLAN for Multimedia Traffic. The work is further extended by comparing the performance of MLAN with FDDI and Fast Ethernet LANs. Simulation results show that MLAN protocol has some potential to operate as a Multimedia LAN. However, analysis shows that some modification of the protocol is required to increase the bandwidth utilisation. Ill AAL ABR ARP ATM B-ISDN BT CAC CBR CDV CDVT CLR CSMA/CD DLPI DVI ELAN FDDI GFC HDTV IEEE IETF IP IPX ISO ITU-T JPEG LANE LAN LEC Acronyms A TM Adaptation Layer Available Bit Rate Address Resolution Protocol Asynchronous Transfer Mode Broadband ISDN Burst Tolerance Connection Admission Control Constant Bit Rate Cell Delay Variation Cell Delay Variation Tolerance Cell Loss Ratio Carrier Sense Multiple Access with Collision Detection Data Link Provider Interface Digital Video Interactive Emulated LAN Fiber Distributed Data Interface Generic Flow Control High-Definition TV Institute of Electrical and Electronic Engineers Internet Engineering Task Force Internet Protocol Internet Packet Exchange International Organization for Standardization International Telecommunications Union - Telecommunications Joint Photographic Expert Group LAN Emulation Local-Area Network LAN Emulation Client lV LECS LES LE ARP LLC LUNI MAC MCDV MCLR MCR MCTD MII MPEG NDIS NetBIOS NNI NTSC ODI P-NNI P-UNI PCI PCR PHY PMI PMD PVC PVP QOS SHD SVC TCP UBR LAN Emulation Configuration Server LAN Emulation Server LAN Emulation ARP Logical Link Control LAN Emulation User to Network Interface Medium Access Control Maximum Cell Delay Variation Maximum Cell Loss Ratio Minimum Cell Rate Maximum Cell Transfer Delay Media-Independent Interface Motion Pictures Expert Group Network Driver Interface Specification Network Basic 1/0 System Network Node Interface National Television Standards Committee Open Data Link Interface Private NNI Private UNI Protocol Control Information Peak Cell Rate Physical layer Physical Medium Independent Physical Mmedium Dependent Permanent Virtual Circuit Permanent Virtual Path Quality Of Service Super High Definition TV Switched Virtual Connection Transmission Control Protocol Unspecified Bit Rate V UNI User-Network Interface VBR Variable Bit Rate VCI Virtual Channel Identifier vc Virtual Circuit VF Variance Factor VPI Virtual Path Identifier VP Virtual Path Vl Table of Contents Acknowledgements Abstract Acronyms Chapter 1 - Introduction 1.0 General Introduction 1.1 Aim of the Research 1.2 Thesis Structure Chapter 2 - Overview of Asynchronous Transfer Mode (ATM) for Multimedia Communication 11 111 IV 1 3 4 2.0 Introduction 5 2.1 Multimedia Communication Applications and their characteristics 5 2.1.1 Characteristics of Video/Image signal 2.1.2 Characteristics of Voice/Speech signal 2. 1.3 Characteristics of Data Traffic 2.1.4 Goals for real-time communication technique 2.2 Architecture of a Generic LAN 2.2.1 Logical Link Control sublayer 2.2.2 Medium Access Control (MAC) sublayer 2.3 Existing LAN Technologies for multimedia communication 2.3.1 Switched Ethernet 2.3.2 Fast Ethernet (l00Base-T) 2.3.3 Fiber Distributed Data Interface (FDDI) 2.4 Evolving LAN Technologies 2.4.1 ISO-Ethernet (ISLAN16-T) 2.4.2 1 00Mbps Demand Priority LAN 2.5 Asynchronous Transfer Mode (ATM) 2.6 ATM System Architecture 2.6.1 The Physical layer 2.6.2 ATM layer Vil 5 11 12 13 14 15 15 16 16 16 17 17 18 19 19 23 24 26 2.6.3 ATM Adaptation layer 27 2.7 Capabilities of ATM 29 2.7.1 Statistical multiplexing 29 2.7.2 Traffic Integration 30 2.7.3 Network simplicity 30 2.8 Disadvantages of ATM 31 2.9 Traffic management in ATM Networks 32 2.9.1 Quality of Service(QoS) and traffic attributes 32 2.9.2 Service categories 34 2.10 Congestion schemes 35 2.11 Credit-based schemes 36 2.12 Rate-based Approach 36 2.12.1 Forward Explicit Congestion Notification Rate Control Scheme 36 2.12.2 Backward Explicit Congestion Notification Rate Control Scheme 37 Chapter 3 - Architecture of High Speed Local Area Networks for Multimedia Communication 3.0 Introduction 3.1 Architecture of ATM LAN 3.1.1 A TM LAN requirements 3.2 ATM LAN Emulation 3.3 Architecture of Multimedia Local Asynchronous Transfer Mode Network 3.3.1 Overview of Multimedia Local ATM Network (MLAN) 3.3.2 Determination and Broadcast ofMLAN parameters 3.3.3 Polling and Bandwidth Reservation 3.3.4 Transmission of Information 3.4 A Comparison of ATM LAN, MLAN, FDDI, Fast Ethernet, Switched Etherent, and ISO-Ethernet Vlll 39 40 42 43 48 48 50 51 54 55 Chapter 4 - Simulation of Multimedia Local Area Network (MLAN) For Voice and Data Traffic 4.0 4.1 4.2 4.3 4.4 4.5 Introduction Data Traffic Generator Voice Traffic Generator Discrete Event Simulation Model SIMSCRIPT II.5 Programming Language 4.4.1 Process Concept 4.4.2 Resource Concept Description of Simulation Model 4.5.1 Process TERMINAL 4.5 .2 Process POLLING 4.5.2.1 Process VRESA 4.5.2.2 Process DRESA 4.5.3 Routine Transmission 4.6 Validation of the Simulation Model 4.6.1 Validation of the Data Traffic Generator 4.6.2 Validation of the Data Traffic Generator 4.7 Validation of the MLAN Protocol Model 4.7.1 Terminal Activity 4.7.2 Average Queue Length 4.7.3 Queuing Delay 4.7.4 Mean Cell Delay 4. 7 .5 Normalized Throughput 4.7.6 Efficiency 57 57 58 60 62 62 63 63 66 66 69 69 72 73 73 73 75 76 76 77 77 77 78 4.8 Discussion of Simulation Results ofMLAN with Data Traffic Generator 79 4.8.1 Terminal Activity 4.8.2 Average Queue Length 4.8.3 Queuing Delay 4.8.4 Mean Cell Delay 4.8.5 Throughput 4.8.6 Efficiency ix 79 84 85 86 87 89 4.9 Simulation ofMLAN with Voice Traffic Generator 4.9.1 Voice Cell Loss 4.10 Performance of MLAN with Voice and Data Traffic 90 90 93 4.11 Simulation of Fiber Distributed Data Interface (FDDI) and Fast Ethernet 95 4.12 Fiber Distributed Data Interface (FDDI) 4.13 The Architecture of FDDI and Operation of its Timers 4.14 COMNET III 4.15 FDDI Network Simulation Model 4.16 Fast Ethernet and its Simulation Model 4.17 A Comparison ofMLAN, FDDI, and Fast Ethernet 4.18 Conclusions Chapter 5 - Conclusions and Future Work 5.0 Conclusions 5 .1 Scope for Future Work References X 96 97 101 101 105 110 112 114 116 117 List of Figures Figure Number 2.1.1 a Structure of a video signal 6 2.1.1 b Digital images and video characterization 7 2.1.3 Multimedia traffic performance requirements 12 2.2 Comparing the layers of the OSI model with the layers and sub-layers of the IEEE/ISO/ANSI LAN archjtecture 14 2.5. l A TM cell structure at (a) user-network interface 21 (b) network-node interface 21 2.5.2 Virtual connections in ATM (a) VP switch 23 (b) VC switch 23 2.6 B-ISDN protocol reference model and the functions of the layer 24 2.6.1 ATM User Network Interface (UNI) 26 2.7.1 Statistical multiplexing gain (a) no statistical multiplexing gain 29 (b) with statistical multiplexing gain 29 3.1 ATM LAN Architecture 40 3.2.l ATM Forum's LAN Emulation components 46 3.2.2 LAN emulation converts packets into cells 47 3.3.1 MLAN Topology 49 3.3.2 Control Cell Format 51 3.3.3 .1 A typical MLAN transmission frame 53 3.3.3.2 Queues after polling for the situation shown in 3.3.3.1 54 Xl 4.2.1 Transformation Techniques for Generating Talkspurt and Silence Length 59 4.2.2 Brady's distribution 59 4 .5a Block diagram of simulation model 64 4 .5b Flow diagram of the Process Timer 65 4.5.1 Flow diagram of the Process Terminal 67 4 .5.2 Flow diagram of the process POLLING 68 4.5.2.1 Flow diagram of the Process VRESA 70 4.5.2.2 Flow diagram for the Process DRESA 71 4.5.3 Flow diagram of the routine Transmission 72 4.6.1 Traffic plot of the data traffic generator at different inter- arrival times 74 4.6.2 Traffic plot of the voice traffic generator 75 4.8.1.la Terminal activity of 48.4% with data terminals 80 4.8.1.lb Terminal activity of 24.2% with data terminals 80 4.8.1 .2 MLAN' s throughput , average queuing delay and mean cell delay with data traffic generator at various terminal activity levels 81 4.8.1.3 Number of terminals supported at various terminal activity levels at a bus speed of 1 00Mbps with 1 km bus length 83 4.8 .2 Average queue length at various bus lengths 84 4.8.3 Average queuing delay of data terminals at various bus lengths 85 4.8.4 Mean cell delay of data terminals at various bus lengths 86 4.8.5.1 Throughput at various bus lengths 87 4.8.5.2 Throughput at various bus speeds with a bus length of 1 km 88 4.9.1 % of voice cell loss at various bus lengths with a bus speed ofl00Mbps 91 Xll 4.10 % of voice cell loss of voice terminals with and without priority 94 4.13a FDDI protocol architecture 97 4.13b FDDI synchronous data negotiation 99 4.15.1 Network simulation model ofFDDI 102 4.15.2 Traffic plot of message generator 103 4.15 .3a Variation of throughput and average packet delay ofFDDI against traffic load with a TTRT of 1 ms 104 4.15.3b Variation of throughput and average packet delay ofFDDI against traffic load with a TTRT of 1.5 ms 104 4 .16.1 Fast Ethernet and 10 Mb/s Ethernet Protocol sub-layers 106 4 .16.2 Network simulation model of Fast Ethernet 108 4.16.3a Variation of throughput and average packet delay of Fast Ethernet against traffic load 109 4.16.3b Collision statistics of Fast Ethernet against traffic load 110 xiii List of Tables Table Number 2.1.1 Multimedia bandwidth requirements 8 2.6 ATM protocol model functions 25 3.4 Comparison of ATM LAN, MLAN, FDDI, Fast Ethernet, Switched Ethernet and !so-Ethernet 56 4.6.2 Duration of Talkspurt and Silence, and Voice activity with different random number generator for a simulation length of 300 seconds with 30 simultaneous voice calls 74 4.7 Simulation parameters for validation using data and voice traffic generators 76 4.8.5a Performance ofMLAN protocol for data traffic at a bus speed of l00Mbps with a terminal activity of 35% 89 4.8 .5b Performance ofMLAN protocol for data traffic at a bus speed of 150Mbps with a terminal activity of 35% 90 4.9.la Performance ofMLAN protocol for voice traffic at a bus speed of l00Mbps 92 4.9.lb Performance ofMLAN protocol for voice traffic at a bus speed ofl50Mbps 92 4.9.lc Number of voice/speech terminals that can be supported at bus speed of 1 00Mbps at various bus lengths with 0% speech cell loss 93 4.9.ld Number of voice/speech terminals that can be supported at bus speed of l 50Mbps at various bus lengths with 0% speech cell loss 93 xiv 4.10 Characteristics of voice and data terminals at a bus length of 3 km with bus speed of 1 00Mbps with and without priority for voice terminals 95 4 .15 Simulation parameters ofFDDI 103 4 .16 Simulation parameters for the simulation model of Fast Ethernet shown in Figure 4.16.1 108 4.17 Comparison ofFDDI, 100 Base-T, and MLAN 111 xv 1.0 General Introduction Chapter 1 Introduction Until mid 80's, network traffic was almost entirely comprised of voice and data traffic. With the advancement of computer technology, there is an increasing demand for multimedia traffic that comprises of audio, video, image, graphics, text and data. Generally, multimedia traffic requires medium to high data transfer rate or bandwidth. Different compression techniques are used to reduce transmission bandwidth requirements. For example, an MPEG-2 session requires a bandwidth between 4-1 OMbps, to transmit audio and video signals, while the projected required bandwidth for HDTV is between 5-30Mbps [l]. The multimedia traffic requirements are low latency, low jitter, lower packet loss etc. Different high speed network structures have been proposed to support multimedia traffic in Wide Area and Local Area Network environments. Broadband Integrated Services Digital Network (B-ISDN) (2] standard has been developed to integrate various types of traffic. It can offer very high data transmission rate using optical fiber link. B-ISDN is a logical extension of the narrow band ISDN. The B-ISDN will be able to integrate all existing network technologies. In addition to that it will be able to support all future teleservices like video-on -demand, video conferencing, high speed data transfer, videophony, home shopping etc. The need for a flexible network and advances in technology and systems led to the definition of the Asynchronous Transfer Mode (A TM) protocol. The ATM [3] protocol is the standard protocol for the B-ISDN and is standardised by the International Telecommunication Union-Telecommunications (ITU-T). ATM is also accepted as the technology to interconnect computers over ATM Local Area Networks by the computer industry in the ATM forum. From the network architecture point of view, computer networks can be classified into three types of networks such as Local Area Networks (LANs), Metropolitan Area Networks (MANs), and Wide Area Networks (W ANs). The classification is done 1 depending upon the distance the corresponding network is designed to span. While LAN s cover an area of a few Km, on the other hand MANs cover an area of up to several tens of Km. W ANs are generally supported by public carrier services which link users separated by geographically wider distances. Use of high-speed data and multimedia applications such as voice, video, graphics etc. have been increasing rapidly in local and wide area network environments. Such integration yields several benefits such as the economy realized by the shared usage of resources. Other benefits of integration are the ease of use of data resources such as file servers for voice applications; and the facilitation of added functionality in data applications, for example, voice annotation of text files and electronic mail [ 4]. As the services like voice, video, fax, graphics etc. are integrated on to the same LAN, the protocols designed primarily for data transmission may not be suitable to meet the requirements of the multimedia traffic [3]. Therefore, significant amount of research has been carried out [5,6,7,8] to integrate data, voice and video traffic on to a Local Area Network .. These new applications continue to make increasing demands on the performance of LANs. LANs are, therefore, required to provide not only high channel throughput, but also satisfy stringent delay requirements. To meet these increasing demands, it is essential that future LANs be capable of operating at much higher data rates achieving high channel efficiencies and lower delay. Operating at data rates ranging from several Mbps to several Gbps, an A TM network with its flexible traffic handling capacity and high data transmission rate could be able to support the multimedia services. With the increasing demand for multimedia services, a shared media ATM LAN (a non­ switch based) spanning relatively shorter distances of a few km and operating in native mode may be a better choice than a switch based ATM LAN for the reasons enumerated below. 2 In a switch based ATM LAN 1. Each terminal in the user group requires a direct full duplex link to the ATM switch 11. All intra-campus traffic of an organization passes through the A TM switch, which need to relay the same message on all links resulting higher resource requirements for the ATM switch. Hence, the cost of the ATM switch, its installation, operation, and maintenance need to be considered. Therefore, ATM LAN emulation is desirable as a backbone network connecting terminals/workstations and traditional local area networks. However, a shared media LAN based on A TM technology could be an alternative low cost choice for certain multimedia applications. 1.1 Aim of the Research The mam objective of this work is to simulate the Multimedia Local Asynchronous Transfer Mode Network (MLAN) protocol [9] for voice and data traffic in order to evaluate the performance of the protocol. Performance of the MLAN is evaluated by the computer simulation technique, using Data and Voice traffic models. The performance of MLAN was compared with two existing high-speed LANs. The existing LANs used for this study are the Fiber Distributed Data Interface (FDDI) [10,11,12,13] and Fast Ethernet (lOObase-T) [7,14]. The FDDI and Fast Ethernet models were simulated using the COMNET III, a communication network simulation package. These high-speed LANs' performances were evaluated in terms of throughput, end-to-end delay, channel efficiency, and ability to integrate different services. Finally, the performance ofMLAN was compared with that of FDDI and Fast Ethernet using the above parameters. 3 1.2 Thesis Structure Chapter 2 presents an overview of ATM for multimedia communication. In this chapter, the multimedia traffic requirements are discussed. Some of the existing network technologies and protocols, which support multimedia applications, are also discussed. A brief introduction of the ATM protocol is given along with the capabilities of A TM to support multimedia applications in the context of a Local Area Network. In addition, traffic management issues in ATM networks are also discussed. In chapter 3, the architecture of a generic LAN, along with the architectures of ATM LAN and MLAN are discussed. The requirements of A TM LAN in the context of multimedia communication are also briefly discussed. The MLAN protocol is discussed in detail and a comparison of the architecture of MLAN is made with that of ATM LAN, FDDI, and Fast Ethernet. In chapter 4, the simulation models of the traffic generators (data and voice) along with the simulation model of MLAN protocol is discussed. The salient features of the discrete event simulation language are explained. An overview of FDDI and Fast Ethernet protocols is also given. The simulation results of FDDI and Fast Ethernet using COMNET III are discussed. Finally, a comparison of MLAN is made with that of FDDI and Fast Ethernet using the simulation results. In chapter 5, the conclusions drawn from the simulation of MLAN is presented along with that ofFDDI and Fast Ethernet. Finally, the scope for future work in MLAN is discussed. 4