Better to know some
... than all
Broadband ISDN and ATM
Asynchronous Transfer Mode (ATM). ATM provides a means for fast switching and transmission at data rates required by B-ISDN services. It relies on the division of information into small fixed-size packets, called cells, and their demand-based transmission over optical fiber - hence the term asynchronous.
Increasing market demand for data rates substantially greater than those supported by ISDN has lead to the notion of Broadband ISDN (B-ISDN). By relying on optical fiber transmission systems, B-ISDN opens the door to a whole range of new applications (e.g., high quality digital audio, real-time video, pay TV, video phone, high speed LAN connection) with emphasis on interactivity and high speed transfer services.
B-ISDN is developed as an evolution of ISDN and hence follows the same principles. This section provides an overview of B-ISDN, its services, and its protocol architecture.
B-ISDN services are clasified into interactive and distribution services. Interactive services involve the bidirectional flow of user information between two subscribers or between a subscriber and a service provider. Interactive services are divided into three subcategories:
* Conversational services involve real-time exchange of information such as sound, video, data, or entire documents. Typical examples include: videotelephony, video-conference, and high speed data transfer. Video-telephony is like the normal voice telephony service but also includes video capture, transmission, and display capabilities. Video-conference provides voice and video communication between two conference rooms or between a number of individuals.
* Messaging services involve non-real-time exchange of information between subscribers in a store-and-forward fashion. Examples include: video mail and document mail. Video mail supports the exchange and storage of moving video images and sound as messages. Document mail allows documents of different types and formats to be exhanged and stored.
* Retrieval services provide subscribers with retrieval access to centrally-stored public information. Examples include: broadband videotex (retrieval of video images/sequences with sound, text, and graphics), video retrieval (subscriber access to video libraries of movies), and retrieval of high-resolution images and documents from various archives and information centres. Distribution services involve the unidirectional flow of user information from a service provider to a subcriber. Distribution services are divided into two subcategories:
* Distribution services without user presentation control involve the central broadcast of information to a large number of subscribers, where subscribers have no control over the presentation of information. Examples include: broadcast of TV programmes, electronic newspapers, and electronic publishing.
* Distribution services with user presentation control are the same as the previous category except that here the information is offered as cyclicallyrepeated frames, thereby enabling the subscriber to control the start and the order of the presentation of frames. Examples include: electronic newspaper and tele-advertising.
B-ISDN User-Network Interface
In addition to the narrowband channels defined for ISDN, B-ISDN supports the following three User-Network Interfaces (UNIs):
* Symmetric full-duplex 155.52 Mbps. This interface provides the basis for interactive services. This will therefore be the most common interface offered. It is commonly referred to as the 150 Mbps interface. The effective payload of this interface is 149.76 Mbps.
* Symmetric full-duplex 622.08 Mbps. This interface is suitable for situations which involve very high traffic volumes. It is commonly referred to as the 600 Mbps interface. The effective payload of this interface is 599.04 Mbps or 600.768 Mbps, depending on how it is structured.
* Asymmetric full-duplex 622.08/155.52 Mbps. This is a hybrid interface which combines the earlier two. It uses bit rates of 622.08 in the network-to-subscriber direction and 155.52 Mbps in the subscriber-to-network direction. It is suitable for situations where the traffic volume in the network-to-subscriber direction is much higher than in the opposite direction.
The functional groupings and reference points for B-ISDN UNI closely follow those of the narrowband, as illustrated by Figure 12.137. Their role remains identical to those of narrowband ISDN. To highlight the broadband nature of the functional groupings and reference points, the letter 'B' is added (e.g., B-TE1). The R reference point is an exception because it may or may not have broadband capabilities. Accordingly, it may support a B-TE2 or TE2.
Asynchronous Transfer Mode
The recommended switching technology for B-ISDN is the Asynchronous Transfer Mode (ATM). Given the high reliability of optical fiber for transmission of digital information, the significant error control overheads involved in earlier protocols (such as X.25) become very questionable. Like frame relay, ATM takes advantage of this increased reliability to improve network performance by providing a highly streamlined protocol stack.
ATM uses two methods to achieve this. First, it transmit all its information in fixed-size small packets, called cells, hence simplifying the processes of packaging and unpackaging user information. Second, unlike X.25 which requires error control and flow control functions to be performed in a link-by-link fashion, it only requires end-to-end support of these functions.
Channels and Paths
The ATM transport network is divided into two layers, both of which are hierarchically organized.
The ATM layer transport functions are divided into virtual channel level and virtual path level. A Virtual Channel (VC) denotes the transport of ATM cells which have the same unique identifier, called the Virtual Channel Identifier (VCI). This identifier is encoded in the cell header. A virtual channel represents the basic means of communication between two end-points, and is analogous to an X.25 virtual circuit.
A Virtual Path (VP) denotes the transport of ATM cells belonging to virtual channels which share a common identifier, called the Virtual Path Identifier (VPI), which is also encoded in the cell header. A virtual path, in other words, is a grouping of virtual channels which connect the same end-points. This two layer approach results in improved network performance. Once a virtual path is setup, the addition/removal of virtual channels is straightforward.
The physical layer transport functions are divided into three levels of functionality. The transmission path connects network elements that assemble and disassemble the transmission system payload. This payload may contain user or signalling information to which transmission overheads are added. The digital section connects network elements (such as switches) that assemble and disassemble continuous bit/byte streams. The regenarator section is simply a portion of a digital section which connects two adjacent repeaters along a transmission path which is otherwise too long to sustain the signal.
Two types of switching are possible: VP switching and VC switching. A VP switch connects a set of incoming VP terminations to a set of outgoing VP terminations. The mapped VPs will have different VPIs, but the VCI will remain unchanged. A VC switch, on the other hand, connects a set of incoming VC terminations to a set of outgoing VC termination. VCIs are not preserved. Because VCs are embedded in VPs, VC switching also involves VP switching.
An ATM cell consists of 53 consecutive octets, of which 48 octents are used to represent user information and the remaining 5 octets form a header for the private use of the ATM protocols.
The GFC field is intended for the user-network interface only. (It does not appear in the header format of cells used for the internal operation of the network. Its 4 bits are instead used by the VPI.) It can be used by applications to provide endto- end flow control. The VPI and VCI fields collectively support the routing of a cell by the network. The PT field denotes the type of information stored in the information field.
The CLP field prioritizes the order in which cells may be discarded. The HEC field is an 8-bit CRC calculated over the remaining 32 bits in the header. It is used for error detection and single-bit error correction.
The ATM layer uses a cell as its basic unit of communication. Unlike the physical layer, at the ATM layer only the logical structure of a cell is of interest. This was described in Section 12.2.2. Cells which are for the use of the physical layer only are distinguished by having the VPI and VCI set to 0, and the least significant bit of the fourth octet in the cell header set to 1. Except for the HEC field which is not used by the ATM layer, the remaining ATM cell fields are separately discussed below.
Generic Flow Control
The GFC field is 4 bits wide and defaults to 0. The GFC field provides a means of exercising UNI-based control over the flow of traffic to counter overload situations. It only applies to traffic generated by the CPE as opposed to the network. The exact GFC details remain largely undefined. The ATM network itself has no provision for flow control comparable to those provided by packet networks. This essentially refelects its streamlined nature.
Virtual Path Identifier
The VPI field is 8 bits wide for the UNI and 12 bits wide for the Network-Network Interface (NNI). It is used to distinguish between VP links multiplexed into the same physical connection. All cells belonging to the same VP are assigned the same VPI. As explained earlier, certain cells may be assigned predetermined VPIs. A VPI is mapped at a VP link termination, such as a switch. A Virtual Path Connection (VPC) is a concatenation of VP links, along which the VCIs may be mapped link-by-link. A VPC preserves the cell sequence for each of the channels it contains.
Virtual Channel Identifier
The VCI field is 16 bits wide and is used to distinguish between VCs sharing the same VP connection. Combined with the VPI, it provides a complete routing identifier. All cells belonging to the same virtual channel are assigned the same VCI. Predetrmined VCIs are used for cells serving special purposes. A VCI is mapped at a VC link termination, such as a switch.
A Virtual Channel Connection (VCC) is a concatenation of VC links. It preserves the cell sequence. At the UNI, a VCC may be established/released using different methods: through user-network signaling, through meta signaling, or as a result of subscription (without signaling).
This field is 2 bits wide and denotes the type of the information stored in the cell. A PT value of 00 is used for cells carrying user information. Other values can be used for representing control information to provide an inband signaling capability.
Cell Loss Priority
This is a single-bit field and hence may assume a value of 0 or 1. When set to 0, it indicates to the network that this cell should not be discarded as long as possible. In a congestion situation, therefore, the network will give preference to the discarding of cells with a CLP of 1. This field may be set by the user or the network.
ATM Adaptation Layer
The AAL supports the use of non-ATM protocols (e.g., LAP-D) for the transfer of information. It does so by mapping information from other protocols (e.g., LAP-D frames) into ATM cells before transmission and their reassembly upon arrival. Four classes of AAL services have been identified on the basis of the timing relation between source and destination, the bit rate, and the connection mode. The AAL provides four protocol types (types 1-4) to, respectively, support these service classes.
Segmentation and Reassembly Sublayer
SAR handles the segmentation of information from CS into ATM cells, and their subsequent reassembly upon arrival.
The Type 1 SAR PDU is intended for transmissions involving a constant bit rate source (see Class A in Figure 12.147). It consists of 8 bits of overhead and a payload of 47 octets. The overhead is comprised of two fields: a 4-bit Sequence Number (SN) and a 4-bit Sequence Number Protection (SNP). The SN is used for the detection of cell sequence violation (e.g., lost or misinserted cells). The SNP serves as a CRC for error detection and single-bit error correction related to the overhead.
The Type 2 SAR PDU (as well as Type 3 and 4) supports transmissions involving a variable bit rate source. It consists of a payload and four overhead fields. The SN is as before. The Information Type (IT) field is used to denote the relationship of the SAR PDU to the CS PDU (i.e., one of: beginning of message, continuation of message, end of message, or single-segment message). The Length Indicator (LI) denotes the actual number of CS PDU octets carried in this SAR PDU. The CRC field protects the SAR PDU against errors. Type 2 is suitable for the transmission of analog data, such as audio and video, using PCM.
The Type 3 SAR PDU consists of 32 bits of overhead and a payload of 44 octets. The Segment Type (ST) field is similar to the IT field of Type 2. The SN, LI, and CRC fields are as before. The remaining 10 bits are reserved for future use.
The Type 4 SAR PDU is very similar to Type 3, except that here a 10-bit Multiplexing Identifier (MID) is used. It supports the multiplexing of multiple CS PDUs over the same connection. Type 3 and 4 are suitable for data transfer applications.
Two modes of service are provided for Type 3 and 4: stream mode and messaging mode. The stream mode service involves the transfer of fixed-size blocks, where each block is mapped to one cell. It is intended for continuous data at low rates. The message mode service is suitable for framed data (e.g., from LAPD). Each frame is mapped to one or more cells.
The convergence sublayer is responsible for a variety of functions (depending on which of the AAL protocol types described above is used), including the following:
* Indication of the information contained by CS PDUs.
* Segmenation and reassembly of higher layer PDUs into CS PDUs.
* Handling of lost of misinserted cells (following their detection by the SAR sublayer).
* Detection of corrupted CS PDUs and their proper handling.
* Explicit time indication through time stamps (as required by some services).
* Clock recovery (as required by some services).