Chapter 6. COMMUNICATION NETWORKS

---

6.1 ATM Networks

Asynchronous Transfer Mode (ATM), also known as cell relay, is a method for information transmission in small fixed-size packets called cells based on asynchronous time-division multiplexing. ATM technology was proposed as the underlying foundation for the Broadband Integrated Services Digital Network (B-ISDN). B-ISDN is an ambitious very high data rate network that will replace the existing telephone system and all specialized networks with a single integrated network for information transfer applications such as video on demand (VoD), broadcast television, and multimedia communication. These lofty goals not withstanding, ATM technology has found an important niche in providing the bandwidth required for the interconnection of existing local area networks (LAN); e.g., Ethernet.

The ATM cells are 53 bytes long of which 5 bytes are devoted to the ATM header and the remaining 48 bytes are used for the payload. These small fixed-sized cells are ideally suited for the hardware implementation of the switching mechanism at very high data rates. The data rates envisioned for ATM are 155.5 Mbps (OC-3), 622 Mbps (OC-12), and 2.5 Gbps (OC-48).

Figure 1: B-ISDN ATM Reference Model

The B-ISDN ATM reference model is shown in Fig. 1. It consists of several layers: physical layer, ATM layer, ATM Adaptation Layer (AAL), and upper layers. This layer can be further divided into the physical medium dependent (PMD) sublayer and the transmission convergence (TC) sublayer. The PMD sublayer provides an interface with the physical medium and is responsible for transmission and synchronization on the physical medium (e.g., SONET or SDH). The TC sublayer converts between the ATM cells and the frames---strings of bits---used by the PMD sublayer. ATM has been designed to be independent of the transmission medium. The data rates specified at the physical layer, however, require category 5 twisted pair or optical fibers.

The ATM layer provides the specification of the cell format and cell transport. The header protocol defined in this layer provides generic flow control, virtual path and channel identification, payload type, cell loss priority, and header error checking. The ATM layer is a connection-oriented protocol that is based on the creation of end-to-end virtual circuits (channels). The ATM layer protocol is unreliable---acknowledgements are not provided---since it was designed for use of real-time traffic such as audio and video over fiber optic networks that are highly reliable. The ATM layer nonetheless provides quality of service (QoS) guarantees in the form of cell loss ratio (CLR), bounds on maximum cell transfer delay (MCTD), cell delay variation (CDV)---known also as delay jitter. This layer also guarantees the preservation of cell order along virtual circuits.

Figure 2: ATM Adaptation Layer (AAL)

The structure of the ATM Adaptation Layer (AAL) is illustrated in Fig. 2. This layer can be decomposed into the segmentation and reassembly sublayer (SAR) and the convergence sublayer (CS). The SAR sublayer converts between packets from the CS sublayer and the cells used by the ATM layer. The CS sublayer provides standard interface and service options to the various applications in the upper layers. This sublayer is also responsible for converting between the message or data streams from the applications and the packets used by the SAR sublayer. The CS sublayer is further divided into the common part convergence sublayer (CPCS) and the service specific convergence sublayer (SSCS).

Initially four service classes were defined for the AAL (Class A--D). This classification has subsequently been modified by the characterization of four protocols: Class A is used to represent real-time (RT) constant bit-rate (CBR) connection-oriented (CO) services handled by AAL-1. This class includes applications such as circuit emulation for uncompressed audio and video transmission. Class B is used to define real-time (RT) variable bit-rate (VBR) connection-oriented (CO) services given by AAL-2. Among the applications considered by this class are compressed audio and video transmission. Although the aim of the AAL-2 protocol is consistent with the focus of this presentation, we shall not discuss it in detail since the AAL-2 standard has not yet been defined. Classes C and D support nonreal-time (NRT) variable bit-rate (VBR) services corresponding to AAL-3/4. Class C is further restricted to nonreal-time (NRT) variable bit-rate (VBR) connection-oriented (CO) services provided by AAL-5. It is expected that this protocol will be used to transport IP packets and interface to ATM networks. A summary of the ATM adaptation layer service classes and protocols is presented in Table 1.

Table 1: ATM Adaptation Layer Service Classes

6.2 Internet (IP Networks)

6.3 Wireless Networks

Wireless networks were until recently primarily devoted to paging as well as real-time speech communications. First generation wireless communication networks were analog systems. The most widely used analog wireless communication network is known as the Advanced Mobile Phone Service (AMPS).[1] The AMPS system is based on frequency-division multiple access (FDMA) and uses 832 30 KHz transmission channels in the range of 824—849 MHz and 832 30 KHz reception channels in the range of 869—894 MHz.

Second generation wireless communication networks are digital systems based on two approaches: time-division multiple access (TDMA) and code-division multiple access (CDMA). Among the most common TDMA wireless communication networks are the IS-54 and IS-136 as well the Global Systems for Mobile communications (GSM). The IS-54 and IS-136 are dual mode (analog and digital) systems that are backward compatible with the AMPS system.[2] In IS-54 and IS-136, the same 30 KHz channels are used to accommodate three simultaneous users (six time slots) for transmission at data rates of approximately 8 Kbps. The GSM system originated in Europe is, on the other hand, a pure digital system based on both FDMA and TDMA. It consists of 50 200 KHz bands in the range of 900 MHz used to support eight separate connections (eight time slots) for transmission at data rates of 13 Kbps.[3]

The second approach to digital wireless communication networks is based on CDMA. The origins of CDMA are based on spread-spectrum methods that data back to secure military communication applications during the Second World War.[4] The CDMA approach uses direct-sequence spread-spectrum (DSSS) which provides for the representation of individual bits by pseudo-random chip sequences. Each station is assigned a unique orthogonal pseudo-random chip sequence. The original bits are recovered by determining the correlation (inner product) of the received signal and the pseudo-random chip sequence corresponding to the desired station. The current CDMA wireless communication network is specified in IS-95.[5] In IS-95 the channel bandwidth of 1.25 MHz is used for transmission at data rates of 8 Kbps or 13 Kbps.

Preliminary plans have proposed for the implementation of the third generation wireless communication networks in the International Mobile Communications-2000 (IMT-2000). The motivation of IMT-2000 is to expand mobile communications to multimedia applications as well as to provide access to existing networks (e.g., ATM and Internet). This is accomplished by providing circuit and packet switched channel data connection as well as larger bandwidth used to support much higher data rates. The focus of IMT-2000 is on the integration of several technologies: CDMA -2000, Wideband CDMA (W-CDMA), Universal Wireless Communications-136 (UWC-136), and Wireless Multimedia and Messaging Services (WIMS).

The CDMA-2000 is designed to be a wideband synchronous inter-cell CDMA based network using the frequency-division duplex (FDD) mode and is backward compatible with the existing CDMA-One (IS-95). The CDMA-2000 channel bandwidth planned for the first phase of the implementation will be restricted to 1.25 MHz and 3.75 MHz for transmission at data rates of up to 1 Mbps. The CDMA-2000 channel bandwidth will be expanded during the second phase of the implementation to also include 7.5 MHz, 11.25 MHz, and 15 MHz for transmission that will support data rates that could possibly exceed 2.4 Mbps.

The W-CDMA is a wideband asynchronous inter-cell CDMA (with some TDMA options) based network that provides for both frequency-division duplex (FDD) and time-division duplex (TDD) operations. The W-CDMA is backward compatible with the existing GSM and provides possible harmonization with WIMS. The W-CDMA channel bandwidth planned for the initial phase of the implementation is 5 MHz for transmission at data rates of up to 480 Kbps. The W-CDMA channel bandwidth planned for a later phase of the implementation will reach 10 MHz and 20 MHz for transmission that will support data rates of up to 2 Mbps.

The UWC-136 is envisioned to be an asynchronous inter-cell TDMA based system that permits both frequency-division duplex (FDD) and time-division duplex (TDD) modes. The UWC-136 is backward compatible with the current IS-136 and provides possible harmonization with GSM. The UWC-136 is a unified representation of IS-136+ and IS-136 High Speed (IS-136 HS). The IS-136+ will rely on the currently available channel bandwidth of 30 KHz, for transmission at data rates of up to 64 Kbps. The IS-136 HS outdoor (mobile) channel bandwidth will be 200 KHz for transmission at data rates of up to 384 Kbps; whereas, the IS-136 HS indoor (immobile) channel bandwidth will be expanded to 1.6 MHz for transmission that will support data rates of up to 2 Mbps.

The WIMS is planned to be a wideband asynchronous inter-cell CDMA based system using the frequency-division duplex (FDD) operation and is compatible with ISDN. The WIMS channel bandwidth scheduled for the first phase of the implementation is 5 MHz for transmission at data rates of 16 Kbps. The WIMS channel bandwidth proposed for the second phase of the implementation will expand to 10 MHz and 20 MHz for transmission that will approach 2.4 Mbps.

The larger bandwidth and significant increase in data rates supported by the various standards in IMT-2000 will facilitate image and video communication over wireless networks. Moreover, the packet switched channel data connection option provided by the various standards in IMT-2000 will allow for the implementation of many of the methods and protocols discussed in the previous sections over wireless communication networks (e.g., RTP).