DATAPRO Data Networking 2720 1 Standards

CCITT Packet Switched Networking Standards X Series

In this report: Synopsis

Packet Assembly! Editor's Note Report Highlights Disassembly ...... 3 In 1984, CCITT published Red Books on A packet switched network permits a user's X.21 Interface wide-ranging topics, including the X.25 data terminal equipment (Le., a PC, host Specifications ...... 6 packet-switching standards. A set of revi­ computer, or terminal) to communicate sions to the X Series, the Blue Books, was with the equipment of other geographically Recommendation published in 1989. Since that time, stan­ dispersed users. Data must be presented X.25 ...... 9 dards ratification and publication have to the network in a prescribed manner, been ongoing, continuous processes. Each however. A packet assembler/disassembler Connections future addition or revision to the X Series (PAD), also referred to as data circuit-ter­ Between Packet Switched standards will be made available, as soon as minating equipment (DTE), serves as a net­ Data Networks '" ...... 17 they are finalized, in individual gray book­ work entry/exit point, packetizing and de­ lets. Since the major building blocks of the packetizing data according to the rules Trends in Packet X standards were completed by 1984, all specified by the X.3, X.28, X.29, X.21, Switching ...... 20 post-1984 technical changes are relatively X.25, and X.75 recommendations of the minor; they are discussed in this report, CCITT. however.

Major developments in packet switching center around the development of ISDN­ related technologies, such as fast packet switching and frame relay, which provide integration of voice, video, and data, and support much higher throughput than tra­ ditional X.25 networks. ISDN's relation­ ship with traditional X.25 packet switching is also discussed.

f -By Martin Dintzis Assistant Editor

@ 1992 McGraw-Hili, Incorporated. Reproduction Prohibited. NOVEMBER 1992 Datapro Information Services Group. Delran NJ 08075 USA 2 2720 CCITT Packet Switched Data Networking Networking Standard. Standards XSerie.

computer manufacturers were reluctant to provide the necessary software to handle the variety of network access protocols. Analysis With the adoption of the X Series Recommendations by the CCITT in 1976, the PDNs could offer a standard network access protocol. The CCITT published revisions to these standards in [984 and 1989. Since that time, the In the early days of packet switching, each Public Data ratification and publication of revisions have become a Network (PDN) defined its own network access protocol, continuous, ongoing process. which permitted an appropriately equipped computer to This report focuses on Recommendations X.3, X.28, communicate with other devices on the network through a and X.29 (informally called the Interactive Terminal In­ physical connection to the PDN. Each of these protocols terface [ITI] standards); X.21; X.2S; and X.7S. used a multiplexing technique that enabled a computer to establish and maintain one or more virtual circuits to other network communicating equipment. No industry standard for packet switching existed, however, and most

Table 1. CCITT Recommendations-X Series

CCITT . Description CCITT Description Recommendation Recommendation X.1 International user classes of service in. X.27 Electrical characteristics for balanced public data networks: class 8 (2400 double-current interchange circuits for bps); class 9 (4800 bps); class 10 data communications equipment (9600 bps); or class 11 (48.000 bps) X.28 DTE/DCE interface for asynchronous X.2 International user faCilities in public device access to the PAD facility of a data networks public data network in the same X.3 Packet assembly/disassembly (PAD) fa- country cility in a public data network; lists op- X.29 Procedure for the exchange of control tions and defaults for interactive information and user data between a asyn chronous terminal connection to packet mode DTE and a PAD facility X.25 packet networks X.32 Procedure for communications between X.4 General structure of signals of Interna- users and packet networks through tional Alphabet No.5 (IA5) code for the switched telephone network and data transmission over public data through circuit switched public data networks (IA5 is described in CCITT networks V.3) X.75 Expanded X.25 recommendation for in- X.20 Interface between data terminal equip- ternetwork communications between ment (DTE) and data circuit-terminat- packet switched networks; interface is ing equipment (DCE) for async defined between two STEs (Signaling trans mission services on public data Terminal Equipments) that are a part networks of ISDEs (International Data Switching X.20 bis V.21-compatibleinterface between DTE Exchanges). with expanded support and DCE for async transmission ser- for wideband links. extended sequenc- vices on public data network ing. and an expanded network utility field for international call X.21 General-purpose interface between establishment DTE and DCE for synchronous opera- tion on public data networks X.92 Hypothetical reference connections for public synchronous data networks X.21 bis For use on public data networks by DTE that are designed to interface to X.95 Network parameters in public data synchronous V-Series modems networks X.24 List of definitions of interchange circuits X.96 Call progress signals in public data between DTE and DCE on public data networks networks X.121 International numbering scheme for X.25 Interface between DTE and DCE for ter- multi network communications contain- minals operating in the packet mode ing a 4-digit DNIX (Data Network Iden- on public data networks tification Code). 3-digit area code. 5- digit host identification. a 0- to 2-digit X.26 Electrical characteristics for unbalanced subaddress double-current interchange circuits for data communications equipment "

NOVE:MSER 1992 @ 1992 McGraw-Hili, Incorporated. ReprOduOtlon Prohibited. Oatapro IhformatiOll Services Group. Delran NJ 08075 USA DATAPRO Data Networking 2720 1 Standards

( CCITT Packet Switched Networking Standards X Series

In thl. report: .ynops.s

Packet Assembly/ Editor's Note Report Highlights Disassembly ...... 3 In 1984, CCITT published Red Books on A packet switched network permits a user's X.21 Interface wide-ranging topics, including the X.25 data terminal equipment (i.e., a PC, host Specifications ...... 6 packet-switching standards. A set of revi­ computer, or terminal) to communicate sions to the X Series, the Blue Books, was with the equipment of other geographically Recommendation published in 1989. Since that time, stan­ dispersed users. Data must be presented X.2S ...... 9 dards ratification and publication have to the network in a prescribed manner, been ongoing, continuous processes. Each however. A packet assembler/disassembler Connections future addition or revision to the X Series (PAD), also referred to as data circuit-ter­ Between Packet Switched standards will be made available, as soon as minating equipment (DTE), serves as a net­ Data Networks ...... 17 they are finalized, in individual gray book­ work entry/exit point, packetizing and de­ lets. Since the major building blocks of the packetizing data according to the rules Trends In Packet X standards were completed by 1984, all specified by the X.3, X.28, X.29, X.21, Switching ...... 20 post-1984 technical changes are relatively X.25, and X.75 recommendations of the minor; they are discussed in this report, CCITT. however.

Major developments in packet switching center around the development of ISDN­ related technologies, such as fast packet switching and frame relay, which provide integration of voice, video, and data, and support much higher throughput than tra­ ditional X.25 networks. ISDN's relation­ ship with traditional X.25 packet switching is also discussed.

( -By Martin Dintzis Assistant Editor

C) 1992 McGraw-Hili. Incorporated. Reproduction Prohibited. AUGUST 1992 Datapro InfonnatJon Services Group. Delran NJ 08075 USA 2 2720 CCITTPacketSwitChed Data Networking Networkl" Standards Standards XSerie.

computer manufacturers were reluctant to provide the necessary software to handle the variety of network access protocols. Analysis With the adoption of the X Series Recommendations by the CCnT in 1976, the PDNs could offer a standard network access protocol. The recommendations are con­ tinually fine-tuned. Revised editions are published at four­ In the early days of packet switching, each Public Data year intervals; the 1989 draft incorporates the latest revi­ Network (PDN) defined its own network access protocol, sions and recommendations. The next published revision which permitted an appropriately equipped computer to will be available in 1992. communicate with other devices on the network through a This report focuses on Recommendations X.3, X.28, physical connection to the PDN. Each of these protocols and X.29 (informally called the Interactive Terminal In­ used a multiplexing technique that enabled a computer to terface [ITI]standards); X.21;X.25; and X.75. establish and maintain one or more virtual circuits to other network communicating equipment. No industry standard for packet switching existed, however, and most

Table 1. CCITT Recommendations-X Series

CCITT Description CCITT Description Recommendation Recommendation X.1 International user classes of service in X.27 Electrical characteristics for balanced public data networks: class 8 (2400 double-current interchange circuits for bps); class 9 (4800 bps); class 10 data communications equipment (9600 bps); or class 11 (48,000 bps) X.28 DTE/DCE interface for asynchronous X.2 International user facilities in public device access to the PAD facility of a data networks public data network in the same country X.3 Packet assembly/disassembly (PAD) fa- cility in a public data network; lists op- X.29 Procedure for the exchange of control tions and defaults for interactive asyn- information and user data between a chronous terminal connection to X.25 packet mode DTE and a PAD facility packet networks X.32 Procedure for communications between X.4 General structure of signals of Interna- users and packet networks through tional Alphabet No. 5 (IA5) code for. the switched telephone network and data transmission over public data through circuit switched public data networks (IA5 is described in CCITT networks V.3) X.75 Expanded X.25 recommendation for in- X.20 Interface between data terminal equip- ternetwork communications between ment (DTE) and data circuit-terminat- packet switched networks; interface is ing equipment (DCE) for async trans- defined between two STEs (Signaling mission services on public data Terminal Equipments) that are a part networks of ISDEs (International Data Switching X.20 bis V.21-compatible interface between DTE Exchanges), with expanded support for wideband links, extended sequenc- and DCE for async transmission ser- ing, and an expanded network utility vices on public data network field for international call X.21 General-purpose interface between establishment DTE and DCE for synchronous opera- X.92 Hypothetical reference connections for tion on public data networks public synchronous data networks X.21 bis For use on public data networks by X.95 Network parameters in public data DTE that are designed to interface to networks synchronous V-Series modems X.96 Call progress signals in public data X.24 List of definitions of interchange circuits networks between DTE and DCE on public data networks X.121 International numbering scheme for X.25 Interface between DTE and DCE for ter- multi network communications contain- minals operating in the packet mode ing a 4-digit DNIX (Data Network Iden- on public data networks tification Code), 3-digit area code, 5- digit host identification, a 0- to 2-digit / ! X.26 Electrical characteristics for unbalanced subaddress double-current interchange circuits for ~ data communications equipment

AUGlJST 1992 @ 1992 McGraw-Hili, InCOtpo,rated, ReprO!luCllon Prohibited. Datapro Information Servlce~: ~rOt.lI!' PlIl~a~N~,Q8~7$,,",S"'

Packet Assembly/Disassembly. The current value (the binary representation of the deci­ mal value) of each PAD parameter delimits the opera­ Recommendations X.3, X.28, and X.29 define the proce­ tional characteristics of the related function. The initial dures by which asynchronous terminals, computers, and value of each parameter is set according to a predeter­ other devices, often referred to as data terminal equipment mined set of values, the initial standard profile. Twenty­ (DTE), communicate with other devices via a packet two PAD parameters have been standardized by the switched network. Packet assemblers/disassemblers, also CCITT. They are as follows: referred to as DTE, commonly serve as network entry/exit points. • PAD Neall using a character-allows an asynchronous X.3 defines the basic and user-selectable functions of a terminal to initiate an escape from the data transfer state packet assembler/disassembler (PAD). It also lists 22 pa­ or the connection-in-progress state in order to send PAD rameters necessary to characterize a specific device (e.g., command signals. This parameter has the following se­ bit rate, the escape character, and flow control technique). lectable values: not possible, possible by character 110 The proper setting of these values enables the PAD to cor­ (DLE), or possible by a user-defined graphics character. rectly interpret the communicating device and vice versa. • Echo-enables characters received from the asynchro­ X.28, a related standard, defines the procedures for nous terminal to be interpreted by the PAD and trans­ character interchange and service initialization, the ex­ mitted back to the asynchronous terminal. Selectable change of control information, and the exchange of user values are no echo (O) and echo (1). data between an asynchronous terminal device and a PAD. • Selection ofdata forwarding characters-allows the asyn­ X.29 defines the procedures for the exchange of PAD con­ chronous terminal to send defined sets of characters, trol information and the manner in which user data is which the PAD recognizes as an indication to complete transferred between a packet mode DTE and a PAD or the packet assembly and to forward a complete packet between two PADs. sequence as defined in X.25. The basic functions of this Recommendation X.3 parameter are encoded and represented by a decimal value. The functions include no data forwarding charac­ CCITT Recommendation X.3, Packet Assembly/Disas­ ter (represented by decimal O); alphanumeric characters sembly Facility in a Public Data Network, outlines the pro­ A-Z, a-z, and 0-9 (decimall); CR (decimal 2); ESC, BEL, cedures for packet assembly/disassembly in asynchronous ENQ, and ACK(decimal4); DEL, CAN, and DC2 (dec­ transmissions. These functions can be programmed and imal 8); ETX and EOT (decimal 16); HT, LF, VT, and built into a microprocessor-based "black box" that is FF (decimal 32); and all other characters in columns 0 placed between the terminal and the X.25 network at ei­ and 1 ofInternational Alphabet No.5 (IA5) not included ther the customer's premises or the entry point of the net­ in the above (decimal 64). work node. The PAD performs a number of functions, some of • Selection ofidle timer delay-permits the selection of the which allow it to be configured, by either an asynchronous duration of a time interval between successive charac­ terminal device or another (remote) PAD, so that its oper­ ters. When data received from the asynchronous termi­ ation is adapted to the asynchronous terminal's character­ nal exceeds this interval, the PAD terminates the assem­ istics. The PAD's basic functions include: bly of a packet and forwards it as defined in the X.25 protocol. • The assembly of characters into packets; • Ancillary control-defines flow control between the • The disassembly of the user data field; PAD and the asynchronous terminal. Decimal 0 repre­ • Virtual call setup, clearing, resetting, and interrupt pro­ sents no use ofX-on (DCI) and X-off (DC3); decimal 1 cedures; represents use of X-onlX-off (data transfer}; and decimal • Generation of service signals; 2 represents the use of X-onlX-off (data transfer and command). • A mechanism for forwarding packets when the proper conditions exist; • Control of PAD service signals-provides the asynchro­ nous terminal with the capability to decide whether and • A mechanism for transmitting data characters, including in what format PAD service signals are transmitted. start, stop, and parity elements; • Selection of operation of the PAD on receipt of the break • A mechanism for handling a break signal from an asyn­ signal-after receiving a break signal from the asynchro­ chronous terminal; nous terminal, the PAD may do nothing, send an inter­ • Editing of PAD command signals; rupt packet to a packet mode DTE or another PAD, reset, or send an indication ofbreak PAD message to a packet • A mechanism for setting and reading the current value of mode DTE or another PAD. PAD parameters; • Discard output-permits a PAD to discard the content of • A mechanism for the selection of a standard profile (op­ user sequences in packets rather than disassembling and tional); transmitting them to the asynchronous terminal. Selec­ • Automatic detection of data rate, code, parity, and oper­ tions include normal data delivery or discard output. ational characteristics (optional); and • Padding after carriage retum-permits the PAD to auto­ • A mechanism for the remote DTE to request a virtual matically insert padding characters in the character call between an asynchronous terminal and another DTE stream sent to the asynchronous terminal after the occur­ ( (optional). rence of a carriage return character. This enables the asynchronous terminal printing device to perform the The PAD's operation depends on the selectable values of carriage return function correctly. A value between 0 and internal variables called PAD parameters. A set of param­ 255 indicates the number of padding characters the PAD eters exists independently for each asynchronous terminal. will generate.

@ 1992 McGraw·HIII,lncorporated. Reproduction Prohibited. AUGUST 1992 Datepro Inforina110n Services Group. DelranNJ 08075 USA 4 2720 CCITT Packet Switched Data Networking Networking Standard. Standards XSerie•

• Line folding-permits the PAD to automatically insert Recommendation X.28 appropriate format effectors in the character stream sent CCITT Recommendation X.28, titled DTE/DCE Inter­ to the asynchronous terminal. No line folding or a prede­ face for a Start/Stop Mode Data Terminal Equipment Ac­ termined maximum number of graphics characters per cessing the Packet Assembly/Disassembly Facility (PAD) in line may be selected. a Public Data Network Situated in the Same Country, de­ • Binary speed-a read-only parameter that neither DTE scribes the interfacing procedures that allow the PAD to be can change. It enables the packet mode DTE to access a connected to an asynchronous terminal. X.28 covers four characteristic (known by the PAD) of the asynchronous areas: terminal device. Speeds from 50 bps to 64K bps are rep- • Procedures to establish an access information path be­ resented. " tween an asynchronous terminal and a PAD; • Flow control of the PAD by the start/stop mode DTE­ • Procedures for character interchange and service initial­ governs flow control between the asynchronous terminal ization between an asynchronous terminal and a PAD; and the PAD. The asynchronous terminal transmits spe­ cial characters to indicate whether it is ready to accept • Procedures for the exchange of control information be­ characters from the PAD. In IA5, these special charac­ tween an asynchronous terminal and a PAD; and ters switch an ancillary transmit device on and off. Dec­ • Procedures for the exchange of user data between an imal 0 represents no use of X-on (DCl) and X-off(DC3); asynchronous terminal and a PAD. decimal 1 represents use ofX-onlX-off. • Line-feed insertion ofter carriage return-permits the Modems standardized for use on public switched or leased PAD to automatically insert a line-feed character in the line facilities establish the procedures for providing an ac­ character stream sent to or received from the asynchro­ cess path (DTE/DCE interface). Procedures for both V and nous terminal or after echo of each carriage return char­ X Series interfaces are defined. acter. This function applies only in the data transfer Transmission speeds up to 1200 bps are specified for state. V-Series interfaces; they are in accordance with either the V.2l, V.22, or V.23 standard, depending on facility type • Line-feed padding-permits the PAD to automatically and speed. The V-Series specifications define the proce­ insert padding characters in the character stream trans­ dures for setting up and disconnecting the access informa­ mitted to the asynchronous terminal after the occurrence tion path by both the DTE and the PAD. of a line-feed character. This enables the asynchronous X-Series interfaces also are used with switched or leased terminal printing mechanism to perform the line-feed line facilities. The physical characteristics for the DTEI operation correctly. This function applies only in the DCE interface are specified in X.20 or X.20 bis. Proce­ data transfer state. dures for setting up and disconnecting the path by both the • Editing-enables character delete, line delete, and line DTE and the PAD are defined. display editing capabilities. During the PAD command X.28 specifies procedures for character interchange and state, the editing function is always available; use or no­ service initialization between an asynchronous terminal nuse of the editing function in the data transfer state is and a PAD. Characters sent and received must conform to selectable. IA5. The PAD transmits and expects to receive only eight­ bit characters. The eighth bit, the last bit preceding the • Character delete, line delete, and line display-all editing stop element, is used for parity checking. functions represented by one user-selectable character X.28 describes the action the PAD takes when the value from IA5. of parameter 21 (X.3, parity treatment) is set to 0, 1,2, or • Editing PAD service signals-enable the asynchronous 3. If parameter 21 is set to 0, the PAD inspects only the terminal to edit PAD service signals for printing devices first seven bits and ignores the eighth bit. When parameter and display terminals; also used for editing via one char­ 21 is set to 1, the PAD treats the eighth bit of the character acter from IA5. Editing is not selectable. as a parity bit and checks this bit against the type of par­ • Echo mask-when echo is enabled, echo mask designates ity-odd, even, space (0), or mark (I)-used between the that selected defined groups of characters sent by the PAD and the asynchronous terminal. If it is set to 2, the asynchronous terminal are not transmitted back. The PAD replaces the eighth bit of the characters to be sent to following may be selected: no echo mask; no echo of CR; the terminal with the bit that corresponds to the type of no echo ofLF; no echo ofVT, HT, and FF; no echo of parity used between the PAD and terminal. When the BEL and BS; no echo of ESC and ENQ; no echo of ACK, value is set to 3, the PAD checks the parity bit for charac­ NAK, STX, SOH, EOT, ETB, and ETX; no echo of edit­ ters received from the asynchronous terminal and gener­ ing characters; or no echo of all other characters. ates the parity bit for characters to be sent to the asynchro­ nous terminal (as in values 1 and 2). • Parity treatment-permits the PAD to check parity in Once the access information path is established, the the datastream from the asynchronous terminal and/or asynchronous terminal and the PAD exchange binary 1 generate parity in the datastream to the asynchronous across the interface. This places the interface in the active terminal. No parity checking or generation, parity check­ link state (state 1). When the interface is in the active link ing, or parity generation are selectable. state, the DTE transmits a sequence of characters that in­ • Page wait-allows the PAD to suspend transmission of dicates service request (state 2) and initializes the PAD. additional characters to the asynchronous terminal after The service request permits the PAD to detect the data the PAD has transmitted a specified number of line-feed rate, code, and parity used by the asynchronous terminal characters. (DTE) and to select the initial profile of the PAD. The ser­ vice request may be bypassed, if the terminal is connected to the PAD via a leased line and the PAD knows the speed, code, and initial profile of the terminal or ifa default value

AUGUST 1992 @ 1992 McGraw-Hili, Incorporated. Reproduction Prohlbijed. Datapro Information Services Group. Delran NJ 08075 USA Data Networking cCln Packet SwHched 2720 Networking Standard. s XSerie. Standards

is used. After the request service signal is transmitted, the been received to fill a packet, when the maximum assem­ DTE transmits binary I, which places the interface in the bly timer delay period has elapsed, when a data forwarding DTE waiting state (state 3A). If parameter 6 (X.3, control character is transmitted, or when a break signal is trans­ of PAD service signals) is set to 0, the interface immedi­ mitted (parameter 7 is set to other than 0). ately enters the PAD waiting state (state 5) after receipt of service request. If parameter 6 is set to other than 0, the Recommendation X.29 PAD transmits the PAD identification PAD service signal CCITT Recommendation X.29, titled Procedures for the (indicates PAD and port identity; is network dependent), Exchange ofControl Information and User Data Between a and the interface enters the service ready state (state 4). Packet Assembly/Disassembly Facility (PAD) and a Packet The DTE then transmits a selection PAD command signal Mode DTE or Another PAD, provides the final step. X.29 (state 6), and the PAD transmits an acknowledgment PAD describes the interfacing procedures that allow the PAD to service signal, followed by binary I, which places the inter­ communicate with the X.25 network. It defines the proce­ face in the connection-in-progress state (state 7). dures for the exchange of PAD control information and If parameter 6 is 0, the PAD will not transmit PAD the manner in which user data is transferred between a service signals. In this case, the interface is placed in the packet mode DTE and a PAD or between two PADs. connection-in-progress state after receipt of a valid selec­ Recommendation X.29 specifies that control informa­ tion PAD command signal. tion and user data are exchanged between a PAD and a Once the DTE receives the PAD service signal (state 8) packet mode DTE or between PADs using the data fields or a sequence of signals in response to a PAD command described in X.2S. Interface characteristics-mechanical, signal, the interface is placed in either the PAD waiting electrical, functional, and procedural-are also defined as state (state 5) or the data transfer state (state 9). inX.25. Afault condition exists if a valid service request signal is X.29 specifies that the call user data field of an incom­ not received by the PAD within a selectable number of ing call or call request packet going to/from the PAD or the seconds after the transmission of binary 1. If a fault condi­ packet mode DTE must consist of protocol identifier and tion occurs, the PAD performs clearing by disconnecting call data fields. A call request packet need not contain a the access information path. call user data field to be accepted by the PAD. If the call The procedures for the exchange of control information user data field is present, the PAD transmits it, unchanged, between an asynchronous terminal and a PAD include to its destination. PAD command signals, PAD service signals, break signals, A call data field's octets consist of user characters sent and prompt PAD service signals. PAD command signals from the DTE to the PAD during the call establishment flow from the DTE to the PAD; they set up and clear a phase. This field is limited to 12 octets. The octet's bits are virtual call, select a standard profile (PAD parameters) numbered 8 to I; bit 1, the low order bit, is transmitted that is either CCITT or network defined, request current first. values of PAD parameters, send an interrupt requesting Bits 8, 7, 6, and 5 of octet 1 of a user data field of com­ circuit status, and reset a virtual call. PAD service signals plete packet sequences are the control identifier field. This flow from the PAD to the DTE; they transmit call progress field, which consists of four octets, identifies the facility to signals, acknowledge PAD command signals, and transmit be controlled. The control identifier field coding for mes­ operating information of the PAD to the terminal. Either sages to control a PAD for an asynchronous terminal is the PAD or the terminal can transmit the break signal. It 0000. When the control identifier field is set to 0000, bits provides signaling without losing character transparency. 4, 3, 2, and I of octet 1 are defined as the message code The prompt PAD service signal indicates the PAD's readi­ field, which is used to identify specific types of PAD mes­ ness to receive a PAD command signal. sages. The temporary storage of characters in an editing buffer User sequences perform data exchange. They are trans­ provides editing functions in the PAD. These functions ferred in the user data fields of complete packet sequences permit the asynchronous terminal to edit characters input with the Q bit set to O. Only one user sequence exists per to the PAD before the PAD processes them. They include complete packet sequence. The PAD transmits all data character delete, line delete, and line display. Character packets with the D bit set to O. The DTE sends a data delete removes the last character in the editing; line delete packet to the PAD with the D bit set to 1. When the PAD removes the contents of the editing buffer. Line display receives a data packet with the D bit set to 1, it sends the causes the PAD to send a format effector followed by the corresponding acknowledgment. The PAD may reset the contents of the editing buffer to the terminal. virtual call, if it does not support the D bit procedure. Procedures for the exchange of user data between an Control information is exchanged via PAD messages, asynchronous terminal and a PAD apply during the data which contain a control identifier field and a message code transfer state. The values of the parameters set in X.3 de­ field that may be followed by a parameter field. PAD mes­ termine which characters are transmitted during the data sages are transferred in the user data fields of complete transfer state. For example, if parameters I (PAD recall packet sequences with the Q bit set to 1. Only one PAD using a character), 12 (flow control of the PAD by the start/ message exists per complete packet sequence. The PAD stop mode DTE), 15 (editing), and 22 (page wait) are set to sends all data packets with the D bit set to O. The DTE may 0, any character sequence may be transmitted by the asyn­ send data packets to the PAD with both the D bit and the chronous terminal for delivery to the remote DTE during Q bit set to I. When the PAD receives a data packet with the data transfer state. both the Q and D bits set to 1, the corresponding acknowl­ User data is sent to the asynchronous terminal in octets edgment is transmitted. The PAD may reset the virtual call ( (eight-bit characters) at the appropriate transmission rate if it does not support the D bit procedure. (Figure 5 shows for the asynchronous terminal; the start/stop bits are the bit positions for the Q and D bits.) added to the data characters. Octets are assembled into packets (see X.2S) and forwarded when enough data has

@ 1992 McGraw-HIII, Incorporated. Reproduction Prohibited. AUGUST 1992 Datapro Information Services Group. Delran 'NJ 08075 USA 8 2720 CCITT Packet Switched Data Networking Networking Standards Standards X Series

The PAD forwards a data packet when a set, read, or set connector size. Also, X.21 offers a much more sophisti­ and read PAD message is received or when any of the con­ cated level of control over the communications process. ditions listed in X.28 exist (e.g., enough data has been re­ Another important feature of X.21 is its inherent dialing ceived to fill a packet, the maximum assembly timer delay functions, including the provision for reporting the rea­ period has elapsed, or a data forwarding character is trans­ sons why a call was not completed. This eliminates the mitted). The PAD never forwards an empty data packet. need for a separate data call interface, such as RS-232-C's By sending a set, read, or set and read message to the companion RS-366-A, and in switched facilities, results in PAD, one can read and change the current values of PAD improved response times. parameters. Upon receipt of one of these messages, the Another important aspect is its relationship to X.25. As PAD delivers to the DTE any previously received data be­ the packet-switching technique becomes more widely im­ fore it acts on the PAD message. The PAD responds to a plemented, the demand will be greater for equipment to read or set and read PAD message by sending a parameter meet the X.21 standard. Internationally, the combination indication PAD message, which contains a parameter field ofX.21, the ISO HDLC protocol, and X.25 has been used listing parameter references and current values. Set allows to form an effective communications path. Another boost the changing of parameters. for X.21 is IBM's recognition. X.29 also discusses invitation to clear procedures, X.21 has some shortcomings. It does not permit the which are used to request that the virtual call be cleared by transmission of control information during data transfer. the PAD; interrupt and discard procedures, which are used Also, it precludes the insertion of data encryption hard­ to indicate that the asynchronous terminal has requested ware between the DTE and the DCE. Another drawback is that the PAD discard received user sequences; reset proce­ the need to modify the DTE/DCE master/slave protocol dures, as defined in X.25; error handling procedures by the techniques and to supply special crossover cables to facili­ PAD, which define the actions to be taken by the PAD tate DTE-to-DTE or DCE-to-DCE interconnection. when errors are detected; and procedures for inviting the X.21 uses a different interfacing technique than that PAD to reselect the called DTE (optional), which are used which is normally associated with physical-level inter­ by a packet mode DTE to request that the PAD clear the faces. Instead of assigning each function a specific pin on virtual call. the connector (e.g., CCITT V.24 and EIA RS-232-C), X.21 assigns coded character strings to each function. The following is a summary of the X.21 standard, in­ X.21 Interface Specifications cluding the functional descriptions of the interchange cir­ The trends in communications engineering lean toward cuits, phases of operation, electrical characteristics, and all-digital networks and the integration of voice and data. mechanical characteristics. Prospective users of these digital, integrated networks are concerned about an interface that can provide access to Functional Descriptions of Interchange Circuits voice, data, video teleconferencing, and related services. Four types of X.21 interchange circuits are defined: Currently, a wide variety of connectors, electrical stan­ Ground, Data Transfer, Control, and Timing. These cir­ dards, and user procedures for various services and net­ cuits, outlined in Table 2, are described below. works exists-leading to almost insurmountable technical Ground and Common Return Circuits-include two and economical problems. Therefore, it is likely that stan­ types of common return circuits, DTE Common Return dards organizations will develop a universal service access and DCE Common Return, and one ground circuit, Signal interface. Although it would require certain extensions, Ground. X.21 is currently the most likely to become a future stan­ Signal Ground (Circuit G) establishes the common ref­ dard for a universal interface in distributed system imple­ erence potential for unbalanced double-current inter­ mentations. change circuits. If required, it reduces environmental sig­ CCITT Recommendation X.21, Interface Between nal interference. Data Terminal Equipment (DTE) and Data Circuit-Termi­ Lowering signaling rates may require two common re­ nating Equipment (DCE) for Synchronous Operation on turn conductors. In this case, two circuits, DTE Common Public Data Networks, defines the physical characteristics Return (Circuit Ga) and DCE Common Return (Circuit and control procedures for an interface between DTEs and Gb), are necessary. For a further explanation of these cir­ DCEs. cuits, see the Electrical Characteristics section of this re­ X.21 is the designated interface for CCITT Recommen­ port. dation X.25, a packet-switching protocol. X.21 can also be Data Transfer Circuits-include two Transmit and Re­ used in a non-packet switched environment. At least two ceive data transfer circuits. X.21-based public data circuit switched networks are cur­ Transmit (Circuit T) transfers signals from the DTE to rently implemented, one in Scandinavia and one in Japan. the DCE during the data transfer phase. It also transfers The X.21 standard has not gained wide acceptance in call control signals to the DCE during call establishment the United States. The reluctance in the U.S. to embrace and other call control phases. the X.21 standard is due in part to the firm entrenchment Receive (Circuit R) receives signals transmitted by the of RS-232-C. Another factor is the cost of implementing DCE from a remote DTE during the data transfer phase. X.21. Since X.21 transmits and interprets coded character This circuit also transfers call control signals from the strings, more intelligence must be built into the interface, DCE during the call establishment and other call control at a higher cost than traditional pin-per-function inter­ phases. faces. Control Circuits-include Control and Indication cir­ Certain characteristics of X.21 should ensure a more cuits. widespread acceptance in the coming years. One immense Control (Circuit C) transmits signals that control the advantage X.21 has over traditional interfaces is its capa­ DCE for a particular signaling process. The representation bility to assign an almost unlimited number of functions, of this signal requires additional coding of the Transmit because there are no functional boundaries associated with

AUGUST 1992 @ 1992 McGraw-Hili, Inccrporated. Reproduction Prohibited. Oatapro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 7 Networking Standard. XSerie. Standards

Figure 1. ( Quiescent States Petlolllpo.·

n Slale number I Signal on T circuil c Signal on C circuH r Signal on R circuil Signal on I clrcuH

TransHion wHh Indication DCE DCE DCE +of whelher OTE or OCE is responsible for IransHlon

The above diagram indicates transitions that are allowed in X.21 networks. Other transitions are possible and may be allowed in some networks. See Table 3 for a listing ofpossible transitions. circuit, as specified for the procedural characteristics of the eight-bit byte. At all other times within the period of the interface. During the data phase, Circuit C remains in the eight-bit byte, Circuit B remains on. the ON condition. Indication (Circuit /) indicates the call control process Pha.e. of Operation to the DTE. The representation of this signal requires ad­ The X.21 standard defines four phases of operation: the ditional coding of the Receive circuit. When Circuit I is Quiescent Phase, the Call Control Phase, the Data Trans­ on, it signifies that signals on the Receive circuit contain fer Phase, and the Clearing Phase. information from the remote DTE. When Circuit I is off, it Each step of the operational phases places the DTE and signifies a control signaling condition, defined by the Cir­ DCE in a certain state. See Table 3 for a listing of these cuit R bit patterns, as specified by the procedural charac­ states and their associated signals on the interchange cir­ teristics of the interface. cuits. Timing Circuits-includes Signal Element Timing and Quiescent Pluue-the quiescent phase is the period dur­ Byte Timing. ing which the DTE and the DCE signal their capability to Signal Element Timing (Circuit S) provides the DTE enter the call control phase or the data transfer phase. It is with signal element timing information. For this function, characterized by the appearance of basic quiescent signals Circuit S turns on and off for nominally equal periods of from the DTE and DCE. Various combinations of these time. quiescent signals result in different interface states, or qui­ X.21 defines different roles for the DTE and DCE in escent states. regard to signal element timing. During the off-to-on tran­ There are three DTE quiescent signals. DTE Ready in­ sition, the DTE presents a binary signal on Circuit T and a dicates the readiness of the DTE to enter the other opera­ condition on Circuit C. The DCE presents a binary signal tional phases. DTE Uncontrolled Not Ready indicates the on Circuit R and a condition on Circuit I during the off­ DTE is unable to enter certain oper,ational phases, usually to-on transition. The DCE transfers the signal element due to an abnormal condition. DTE Controlled Not Ready timing across the interface as long as the timing source is indicates that although the DTE is operational, it is tempo­ capable of generating this information. rarily unable to accept incoming calls for circuit switched Byte Timing (Circuit B) provides the DTE with eight-bit service. timing information for synchronous transmission. Use of There are two DCE quiescent signals: DCE Ready and this circuit is not mandatory. Circuit B turns off whenever DCE Not Ready. DCE Ready indicates the DCE is ready Circuit S is in the ON condition, indicating the last bit of to enter operational phases. DCE Not Ready indicates that

@ 1992 McGraW-Hili. Incorporated. Reproduction ProhlbHed. AUGUST 1992 Datepro Information Services Group. Delran NJ 08075 USA 2720 CCI" Packet ewltch.d Data Networking N.tworklng etandard. • Standards xe.ri••

no service is available; it is also signaled whenever possible Information state immediately after all call progress sig­ during network fault conditions and during the period nals, if any, are transmitted. Both calling and called line when test loops are activated. identification are optional. Line identification consists of the international data number, as assigned in CCITT There are six quiescent states: Recommendation X.121, International Numbering Plan for Public Data Networks. The DeE transmits Charging • Ready Information during the DCE-Provided Information • DTE Controlled Not Ready, DCE Ready state. It. informs the subscriber of either the monetary • DTE Ready, DCE Not Ready charges for a call, the duration of the call, or the number of units used during the call. • DTE Uncontrolled Not Ready, DCE Not Ready • Connection-In-Progress, indicated by the DCE. • DTE Controlled Not Ready, DCE Not Ready • Ready for Data, transmitted by the DCE when the con­ • DTE Uncontrolled Not Ready, DCE Ready nection is available for data transfer between DTEs.

See Figure 1 for a diagram of the quiescent states and the An unsuccessful call occurs when a required connection transitions that are allowed between these states. cannot be established. In this case, the DCE indicates the Odl Control Phase-the call control phase for circuit failure and its reason to the calling DTE through a call switched service contains many elements and procedures. progress signal. Characters used for call control are selected from 1A5, a A call collision can occur in one of two ways: a DTE seven-bit plus parity international code outlined in CCITT detects a call collision when it receives Incoming Call in Recommendation Y.3. Each call control sequence to and response to Call Request. A DCE detects a call collision from the DCE is preceded by two or more continuous SYN when it receives Call Request in response to Incoming characters. For error checking of call control characters, Call. When the DCE detects a call collision, it will indicate odd parity is specified. Proceed to select and cancel the incoming call. The following elements of the call control procedure are The DTE indicates a request for a direct call by signal­ outlined in X.21: events of call control procedures, unsuc­ ing DTE Waiting after receiving the Proceed to Select sig­ cessful call, call collision, direct call, and facility registra­ nal. Ifnecessary, the DTE may choose an addressed call by tion/cancellation procedure. These elements are summa­ presenting the correct Selection signal. rized below. The facility registration/cancellation procedure is op­ The events of the call control procedures include the tional. A facility registration/cancellation signal consists of following: up to four elements in order: facility request code, indica­ • Call Request, signaled by the DTE to indicate a request tor, registration parameter, and address signal. Not all of fora call. these elements are required in the facility registration/can­ cellation signal. Also, a number of these signals may be used when the network is prepared to • Proceed to Select, linked to form a block. In response to acceptance or rejec­ receive selection information. It is transmitted by the tion of the facility registration/cancellation action, the net­ DCE to the DTE within three seconds of the call request signal. . work provides the appropriate Call Progress Signal. Data TrtlIIS/er Phase-when the DTE is in the data • Selection Signal Sequence, transmitted by the DTE. A transfer phase, any bit sequence may be transmitted. X.21 selection sequence consists of a facility request block, an defines the data transfer phase for three types of connec­ address block, a facility request block followed by an ad­ tions: switched; leased, point to point; and leased, central­ dress block, or a facility registration/cancellation block. ized multipoint. A facility request block comprises one or more facility For operation over switched facilities, the DTE may request signals, which consist of a facility request code send bits to a corresponding DTE after receiving the containing one or more facility parameters. An address Ready for Data signal. During data transfer, control and block contains one or more address signals. Address sig­ interchange circuits are in the ON condition, and data is nals consist of either a full address signal or an abbrevi­ transmitted over the transmit and receive circuits. Data ated address signal. transfer may be terminated by clearing, which is defined • DTE Waiting. below. Two basic signals are used for operation over leased, • Incoming Call, indicated by the DCE. In response, the point-ta-point facilities. Send Data transmits data by the DTE signals Clear Request, DTE Uncontrolled Not DTE on Circuit T; the remote DTE's Receive Data signal Ready, or DTE Controlled Not Ready. receives data over Circuit R. To terminate the data trans­ • DCE Waiting. fer, the DTE signal places its transmit circuit in the binary I condition. The DCE indicates termination of data trans­ • Call Progress Signal Sequence is transmitted by the DCE fer by placing its receive circuit in the binary I condition, to the calling DTE to indicate that circumstances have its control circuit in the OFF condition, and its indications arisen to prevent the connection from being established, circuit in the OFF condition. to report the progress made toward establishing the call, Both the central and remote DTEs use the Send Data or to signal that problems have been detected and that and Receive Data signals for operation over leased, mul­ the call needs to be cleared and reset. tipoint facilities. The central DTE delivers data transmit­ • DCE-Provided In/ormation Sequence, transmitted from ted to all remote DTEs; remote DTEs (one at a time) trans­ the DCE to the calling DTE. It consists of DCE-provided mit data to the central DTE. A remote DTE may send data information blocks, such as line identification and to the central DTE while the central DTE is sending to all charging information. Line Identification is transmitted remote DTEs. by the DCE to the calling DTE during the DCE-Provided

AUGUST 1992 @ 1992 McGraw-Hili. Incorporated. Reproduction Prohibited. . Datepro Information Services Group. Dalran NJ 08075 USA Data Networking CCITT Pecke' Switched 2720 Networking Standerd. xSeri•• Standards •

ISO 4903 assigns connector pin numbers to a 15-pin Table 2. CCITT X.21 Interchange interface between DTE and DCE equipment. Table 4 pre­ ( Circuits sents a chart of these X.21 pin assignments as they relate to the X.26 and X.27 standards. Direction X.21 Bis Inter­ Name to DCE from DCE Circuit Although X.21 is the specified interface for X.25, alterna­ change Type tive interfaces also exist. One of these is X.21 bis. Circuit CCITT Recommendation X.21 bis, the physical and G Signal functional equivalent to CCITT V.24, defines 25 inter­ ground or change circuits between DTEs and DCEs. CCITT V.24 is common compatible with EIA Standard RS-232-C. The X.21 bis return recommendation, accepted as the interim interface for X.25, will be gradually replaced by X.21 as more equip­ Ga DTE com- X Ground man return ment is manufactured to meet X.21 specifications. Gb DCE com- X man return Recommendation X.25 T Transmit X Data The development of Recommendation X.25 was stimu­ Transfer lated by the need for a standard interface between the R Receive X packet-switching networks already developed or being de­ veloped by many industrial nations and by the require­ c Control X ment that no terminal equipment be denied access to Indication X Control packet switched services. X.25 is a dynamic standard with many variations in the s Signal ele- X ment timing U.S. and abroad. Currently, X.25-based packet switched networks exist in Australia, Austria, Belgium, Canada, B Byte timing X Timing France, Ireland, Germany, Hong Kong, Italy, Japan, Mex­ ico, the Netherlands, Portugal, Singapore, the Soviet Clearing Phase-either the DTE or the DCE may ini­ Union, South Africa, Spain, Switzerland, the United King­ tiate clearing. The DTE indicates its desire to enter the dom, and the United States. Since X.25 is a dynamic stan­ clearing phase by transmitting DTE Clear Request. The dard with many extensions and optional features, these DCE responds by signaling DCE Clear Confirmation, fol­ networks are not totally compatible with one another. lowed by DCE Ready. Those located in Europe have the highest level of mutual Clearing by the DCE takes place when it transmits DCE compatibility. Clear Indication. The DTE responds with the DTE Clear Since the establishment of X.25, additional user-level Confirmation signal, followed by the DCE signaling DCE protocols have been developed. These protocols provide Ready. the interfaces between different types of terminals and the X.25 interface. X.3, X.28, and X.29, informally called the Electrical Characteristics Interactive Terminal Interface (ITI), were the first of the X.21 uses two types of electrical characteristics, each for protocols to interface to X.25. They relate to the support of different system requirements. asynchronous, low-speed terminals by packet switched For synchronous operation at 9600 bps and below, the networks. These are logical complements to X.25 because interchange circuits at the DeE side of the interface must they permit specific sets of terminals to interface to the comply with CCITT Recommendation X.27. The DTE packet networks using the X.25 interface. side can comply with either X.27 or another CCITT Rec­ The X.25 interface standard provides for the connec­ ommendation, X.26. For synchronous operation at signal­ tion of terminals and computers to public packet-switch­ ing rates above 9600 bps, interchange circuits at both the ing networks. X.25 outlines three layers of operation: the DTE and DCE sides of the interchange circuits must com­ Physical Layer, the Link Layer, and the Packet Layer. ply with X.27. These layers parallel the bottom three layers of the ISO X.26 is defined in CCITT Standard V.lO. It describes Reference Model for Open Systems Interconnection. The the electrical characteristics for unbalanced interchange Physical Layer calls for CCITT X.21 as the physical and circuits. X.26 calls for both a DTE and DCE common electrical interface but accepts X.21 bis, a functional grounding arrangement. The maximum suggested cable equivalent of RS-232-C, as an interim standard. The Link length is 1,000 meters, and the maximum data rate is lOOK Layer uses the procedures of the HDLC protocol standard. bps. The Packet Layer defines procedures for constructing and X.27 is defined by CCITT Standard V.l1, which de­ controlling a data packet. scribes the electrical characteristics for balanced opera­ The 1984 revision of Recommendation X.25 added tion. Maximum suggested cable length is 1,000 meters, specifications for X.21 access and expanded the potential and the maximum data rate is 10M bps. of packet operations, allowing users to actively gain access to the X.25 port, identify themselves, and validate their Mechanical Characteristics connection through passwords. This change reoriented the The mechanical characteristics for X.21 are outlined in the X.25 standard toward switched access through both X.21 ( ISO Standard 4903, approved by the International Organi­ facilities and the public telephone network. It now sup­ zation for Standardization (ISO). The standard, entitled ports X.32 with regard to the public switched telephone Data Communication-I5-pin DTEIDCE Interface Con­ network or a circuit switched public data network, dial-in nector and Pin Assignments, was published in June 1980.

C> 1992 McGraw-HIli, Inccrporated. Reproduction Prohibited. AUGUST 1992 Datepro Information Services Group. Delran NJ 08075 USA 10 2720 eeln Pacleet Switched Data Networking Networldn. Stancill... Standards XS.rI..

Table 3. X.21 States: Names, Signalling, and Transitions ',,-- Signals on T,C and R,I Circuits Transitions* Phase ot DTE DCE State Opera- to state to stete Number Stete Name tion T C R I number number 1 Ready Q 1 OFF 1 OFF 2,13S,14;24 8,13R,18 2 Call request CC 0 ON 1 OFF 3,15 3 Proceed-to-select CC 0 ON + OFF 4,15 4 Selection signal CC IA5 ON + OFF 5

5 DTE Waiting CC 1 ON + OFF 6A,11,12 6A DCE Waiting CC 1 ON SYN OFF 7,10,11,12 6B DCE Waiting CC 1 ON 8YN OFF 10bls, 11,12 7 Call progress signal CC 1 ON IA5 OFF 6A,10,11,12 8 Incoming call CC 1 OFF BEL OFF 15,9 9 Call accepted CC 1 ON BEL OFF 6B,11,12 10 DCE provided information CC 1 ON IA5 OFF 6A,11,12 10 bls DCE provided information CC 1 ON IA5 OFF 6B,11,12 11 Connection in progress CC 1 ON 1 OFF 12 12 Ready for data CC 1 ON 1 ON 13 13 Data transfer DT D ON D ON 13R 13S,DCE not ready

13R Receive data DT 1 OFF D ON 13 1 13S Send data DT D ON 1 OFF 7 13 14 DTE Controlled not ready, DCE ready Q 01 OFF 1 OFF 1,24 23 15 Call collision CC 0 ON BEL OFF 3 16 DTE Clear request C 0 OFF X X 17 (see Note) 17 DCE Clear confirmation C 0 OFF 0 OFF 21 18 DTE Ready, DCE Not ready Q 1 OFF 0 OFF 22 1 DCE Not ready Q D ON 0 OFF 1,13,138 19 DCE Clear indication C X X 0 OFF 20 (see Note)

20 DTE Clear confirmation C 0 OFF 0 OFF 21 21 DCE Ready C 0 OFF 1 OFF 22 DTE Uncontrolled not ready, DCE not ready Q 0 OFF 0 OFF 18 24

23 DTE Controlled not ready, DCE not ready Q 01 OFF 0 OFF 18,22 14 24 DTE Uncontrolled not ready, DCE ready Q 0 OFF OFF 22

Any state (see Note) X X X X 16 19 • All other transitions are considered invalid. Note: DCE Clear Indication or DTE Clear request may be entered from any state except Ready.

Key to Table: a and 1-steady binary conditions 01-Alternate binary 0 and binary 1 Q-Qulescent Phase X-Any value CC-Cail Control Phase OFF-Continuous off (binary 1) DT-Data Transfer Phase ON-Continuous on (binary 0) T-Transmit interchange circuit IA5-Characters from CCITT Alphabet #5 C-Controllnterchange circuit +-:-IA5 character 2/11 R-Receive interchange circuit BEL-IA5 character 0/7 I-Indication interchange circuit SYN-IA5 character 1/6

AUGUST 1992 @ 1992 McGraw-Hili. Inco~ated. Reproduction Prohibited. Datepro InfOrmation SeMces Group. Delran NJ. Q8075 USA

------Data Networking CCITT Packet Switched 2720 11 Networking Standard. X Serle. Standards

and dial-out access, backup for leased line connections, intelligent terminals (multiplexed to avoid the use of mul­ long-distance access to the network, and teletex. tiple lines) that transmit data over the packet network to a Datagram was deleted in 1984, while the following remotely located host. It can also be a processor acting on packet-level services were made available as options: calls received from multiple locations that communicate • Registered Private Operating Agent (RPOA) Selection per­ over the packet network. mits the use of one or more networks to route a call to its Regardless of the device or application, all DTEs destination. If the user selects only one network, either present standard formatted data and control information the basic or extended format of the RPOA Selection can to the DCE over standard communications facilities. De­ be used; if more than one network is chosen, the ex­ vices operate over the network in the virtual circuit mode. tended format is used. Essentially, the user causes the network to establish a logi­ cal circuit connection with the receiving station for the • Call Redirection permits the rerouting of calls if the first transmission of multiple contiguous packets. (The actual tried route fails. physical circuit over which individual packets are trans­ • Call Redirection Notification informs the recipient of the mitted can vary during a session, but the logical circuit forwarded call that the call has been redirected. ~nsures presentation of each packet to the receiving station m the proper order.) Information delivery is so rapid that • Called Line Address Modified NotifICation tells the caller, the user appears to have an end-to-end, dedicated channel. within a call confirmation packet, that the call has been Users who do not support X.25 can access the public redirected. data network for packet data transmission; however, they • Hunt Group distributes incoming calls that have an ad­ cannot transmit directly to a DCE or network node. They dress associated with the hunt group. must transmit to a special network-operated PAD, dis­ • Charging Information gives the caller information on cussed earlier in this report. Terminal transmissions are time and charges and requires a new field in the call­ stored in a buffer at the PAD. There they are assembled clearing packet format. into packets and sent to the device at the other end of the virtual circuit. Packets arriving at the PAD from the net­ • Local Charging Prevention is a security facility that pre­ work are reassembled into the appropriate format before vents reverse or third-party call charges. they are sent on to the terminal. The PAD is programmed • Network User Identification accommodates user ID, bill­ and configured to interface properly with the protocol and ing, and on-line facilities registration. This permits users physical characteristics of the user's device. Data pre­ to communicate directly with the packet data network to sented to the PAD is reformatted into X.25 format and change the parameters of their subscriptions. forwarded to the DCE. Recommendation X.25 is divided into those specifica­ The packet level is an octet-oriented (eight bits per octet) tions required for a device or network to comply and those structure. Packet sizes can vary from 1,024 to 2,048 octets, that are optional. Approximately one third are required; but only within a network. Network-to-network exchange the remaining two thirds of the specifications are optional. is limited to 128 octets. Closed user group facilities can The excess throughput capacity inherent in the X.25 accommodate very large private-packet networks, al­ standard allows for future network growth and technologi­ though the number of closed user groups to which a DTE cal progress. For example, a single X.25 interface can the­ can belong is network dependent. oretically handle 4,095 virtual channels, packet sizes up to Link-level changes implemented in 1984 created a clear 2048 bytes each, and packet sequencing up to 128 packets separation between the Link Access Procedure (LAP) and per logical channel. Most network suppliers' nodal proces­ the Link Access Procedure Balanced (LAPB). Multilink sors are too small to handle this much traffic through a procedures for a single interface were also implemented. single interface. Therefore, in practice, the support offered LAPB underwent an alignment with the single-link proce­ over each interface is limited to the current capacity of the dure of X.75. An extended numbering option (modulo network's access node. 128) was added to LAPB to enable the sequencing of 127 Functional Layers frames and the use of satellite facilities. In addition, the 1984 revisions to X.25 refined the procedure for the imple­ Recommendation X.25 defines three functional layers: the mentation of the D-bit, polished technical accuracy, and Physical Layer (Levell), the Link Layer (Level 2), and the defined the rules for new fields and formats. Packet Layer (Level 3). These are consistent with the first three layers of the ISO Reference Model for Open Systems X.25 Communications Interconnection (OSI). (Although OSI labels its third layer X.25 is titled Interface Between Data Terminal Equipment the Network Layer, its function parallels that of the Packet (DTE) and Data Circuit-Terminating Equipment (DCE) Layer. Both provide the means to establish, maintain, and for Terminals Operating in the Packet Mode on Public Data terminate connections.) Networks. It provides a precise set of procedures for com­ Levell: Physical Level-outlines the physical, func­ munications between DTE and DCE for terminal equip­ tional, and electrical characteristics of the DTElDCE in­ ment operating in a packet environment. The DCE in this terface. X.21 uses the transmission of coded character case is a node processor that serves as an entry/exit point strings across a I5-pin connector to define standard inter­ on the packet network side of the user/network interface. face functions, e.g., Transmit Data. Its specifications are The Data Terminal Equipment (DTE) is a programma­ defined in CCITT Recommendation X.2I. ble device on the user side of the user/network interface. Although X.2I is the recommended physical-level in­ ( The DTE is located at the user site when the on-site equip­ terface for X.25, the availability of data communications ment supports X.25; at such installations, tbe DTE can be equipment with X.21 capabilities is limited, especially in either a computer, a front-end processor, or an intelligent the United States. As a result, the CCITT has accepted an terminal, as shown in Figure 2. The DTE can be a group of interim recommendation, X.21 bis, as the X.25 physical

@ 1992 McGraw·HiII,lncorporated.Reproduction Prohibited. AUGUST 1992 Datapro Information Services Group_ Delran NJ 08075 USA 12 2720 CCI" Packet Switched Data Networking Networkl... alandard. Standards X ......

Figure 2. . A Stutlple U,er Termi"al eo"/iguratio,, lor Operati",o" a Pubik DIIta Network (PDN) i" Paclut Mode, witi X.2S "' tie l"terface to tie Network

User Location Public PlJCIcet-8wltchlng Netwotk

VIrtual Circuit DTE DCE Computer with X.25 Level 2 I .- Modem I .- and Level 3 ./ Modem Software - I I - Node Processor I I with K25 Level 2 DTE I I I and I I Level 3 Front- Software Host .- .- Computer End Modem I ./ I Modem Processor - - I J X.25 Level 2 and 3 Software I

DTE I Intelligent Terminal whh I X.25 Level 2 .- and Level 3 -I I Software I I I I I I .- ~ /' I Modem I I i . - OTE I iii I ~ I Intelligent Terminal wlh I I X.25 Level 2 .- ~ and Level 3 - I ~ Software r--

• Signifies an interface that must conform to X.25 Level 2 Physical Interface Standards. interface. X.21 bis is functionally equivalent to EIA RS- (ARM). Due to some incompatibilities with some ele­ 232-C, which assigns a separate function to each pin across ments of procedures, Level 2 of the X.25 Recommenda­ a 25-pin connector. tion was revised to incorporate LAPB. Similar to the Asyn­ Level 2: Linlc LeJlel-describes the link access proce­ chronous Balanced Mode (ABM) ofHDLC, it provides for dures used for data interchange between a DCE and a a "balanced" configuration; that is, each side of the link DTE. In the CCITI Recommendation X.21, International consists of a combined primary/secondary station. The User Classes o/Service in Public Data Networks, X.25 spec­ Multilink Procedure is used for data transmission on one ifies that the DTE and DCE must operate in the following or more single links. Each link conforms to the X.25-de­ user classes of service: class 8 (2400 bps), class 9 (4800 fined frame structure and to the elements of procedure de­ bps), class 10 (9600 bps), or class 11 (48K bps). The proce­ scribed in LAPB. dure calls for a full-duplex facility so that the DTE and Software in both DTE and DCE performs Level 2 pro­ DCE can conduct two-way, simultaneous, independent cessing. This software appends control information onto transmissions. The procedures use the principles and ter­ packets that are ready for transmission, maintains control minology of the High-level Data Link Control (HDLC) of the transmissions, performs transmission error check­ protocol specified by the International Organization for ing, and strips a successfully received frame down to a Standardization. packet. The packet consists of packet control information Level 2 comprises three procedures: the Link Access and (optionally) user data; packets are discussed in detail Procedure (LAP), the Link Access Procedure Balanced later in this report. (LAPB), and the Multilink Procedure (MLP). The original HDLC specifies certain control fields that must be ap­ X.25 data link control procedure embraced LAP, which is pended to both ends of a packet, resulting ina transmis­ similar to the HDLC Asynchronous Response Mode sion format called a frame. Appended in front of a packet are a beginning Flag field, an Address field, and a Control

AUGUST 1992 @ 1992 McGraw-HIH. IlKXII'PCN"ated. Reproduction Prohibited. Datapro Information Services Group. Delran NJ 08075 USA Data Networking cern Packet SwItched 2720 13 Networking .tII ...... X ...... Standards

Figure 3. Frtune Fornum for PQCket Mode TrtUllmulion

~I"~------Frarne------~"~I

...
..... ----packet----II ..~I

Packet Field Control Flag Address Control User Data Sequence Flag Field Intornation Check

Fields Generated by:

Application Program

Level 3 Software

Level2 Software Size, Field bits Description

8 Valueis01111110. Flag 8 Value for DTE to DCE command frame is "10000000·. Address Value for DTE to DCE response frame is ''11000000·.

Value for DCE to DTE command frame is ''11000000". Value for DCE to DTE response frame is ''10000000".

8 I Frame: Control Bit 1 is "0". Bits 2, 3, 4, are N(S) (transmitter send sequence count). Bit 5 is P/F (PoIVFinal) bit. Bits 6,7,8 are N(R) (transmitter receive sequence count).

S Frame: Bit 1 is "1". Bit 2 is "0". Bits 3, 4 identity Supervisory Command. Bit 5 is P/F (PoIVFinal) bit. Bits 6,7,8 are N(R) (transmitter receiver sequence count).

U Frame: Bit 1 is 1. Bit 2 is 1. Bits 3, 4 are first part of Unnumbered Frame Command identifier. Bit 5 is P/F (PoIVFinal) bit. Bits 6, 7,8 are second part of Unnumbered Frame Command identifier.

Packet 24 Control information. Control 1,024 1,024 data bits are normal maximum; exceptional maximum is 2,040 data bits. User Data 16 Check bits for user data. Field Sequence Check 8 Value is "01111110·. Flag

field. Appended behind the packet are a Frame Check Se­ for which the command is intended in command frames or quence field and a closing Flag field. Figure 3 gives the size to identify the station sending the response in response and description of each field. The receiving device uses the frames. two Flag fields to ascertain the beginning and ending of a The Control field identifies the type of frame and sup­ ( frame. A single Flag field can be used as the closing flag for plies control information pertinent to that type offrame. A one frame and the opening flag of the next frame. The Ad­ frame can be either an Information Frame (I-Frame), a Su­ dress field, under HDLC, is used to identify the station(s) pervisory Frame (S-Frame), or an Unnumbered Frame (U­ Frame). See Table 5 for the encoding of this field.

@ 1992 McGraW-Hili, Incorporated. Reproduction Prohibited. AUGUST 1992 Oatapro Information Sarvices Group. Delran NJ 08075 USA t4 2720 oom Packet Switched Data Networking Networking Standards Standards X ..rI ..

Figure 4. X.2S User and Network Software Relationships

i------USer Localion _____+- ____:Communications ____-+ ____ Public Packet- Facility Switching NelWork

DCE DTE NelWork Node Processor

r----.,-----I I I I I I I Comrnunicetions Aec:esa Method I I I Usar r~---I r----I I I Program I US I I LevelS. I Level 2. I~er : ~;;S~t':~~. ~~~~e~:~: I Network I Packet Level I I Frame Level I ..... I Nodes .... _--_. _--_. I I I I I I I I Physicallnlerface I

l<.25 Levell. LPhysicallnterfaee

An Information Frame contains a user packet. The Lne13: Packet Level-describes the packet format and Control field ofthe I-Frame contains the N(S) transmitter control procedures for the exchange of packets between the Send Sequence Count, the N(R) transmitter Receive Se­ DTE and the DCE. In addition to the user data packet quence Count, and the Poll/Final (P/F) bit. N(S) is the se­ format, there are several formats for DTE/DCE adminis­ quence number of this frame sent from this transmitter to trative messages. this receiver. N(R) is the sequence number of the next The software for formatting packets and controlling frame this transmitter expects to get from the receiver packet exchange is the Level 3 software resident in both / when the receiver becomes a transmitter. The PolllFinal the DTE and the DCE. Figure 4 shows the relationship of bit indicates whether a transmission acknowledgment is Level 3 and Level 2 software operating in the DTE and the required or when the final frame in a stream has been DCE. In the DTE, a user application program normally transmitted. presents the data to be transmitted to the operating sys­ The Supervisory Frame transmits one of three supervi­ tem's communications access method. The access method sory commands and cannot contain user data. The Control invokes Level 3 software to append the packet header in­ field of the S-Frame contains the command, the PIF bit, formation, then invokes Level 2 software to create the and an N(R). Table 6 summarizes the commands and re­ frame. Packet header formats are discussed below. sponses possible in the Control field. Once the frame is created, it is ready for transmission. The Unnumbered Frame extends the number of link Upon receiving the frame, the receiving DCE invokes control functions, without incrementing the sequence Level 2 software to perform error detection and sequence counts at either the sending or receiving station. checking and to strip the accepted frame down to a packet. A U-Frame control field contains the Unnumbered The packet is presented to Level 3 software for packet­ command and the PIF bit. One of the U-Frame commands level processing and to prepare the packet for transmission initializes the DTElDCE network to the Asynchronous Re­ to the destination DCE. The physical address of the desti­ sponse Mode (ARM) of operation, which permits both nation DTE, derived from the call initiation, is inserted DTE and DCE to initiate transmission to each other. into the packet during Level 3 processing. At the network The Frame Check Sequence (FCS) field contains a 16- destination DCE, Level 3 software reformats the packet bit check pattern created at framing time by generating control information and presents the resulting packet to check bits based upon the binary value of the data in the Level 2 software; Level 2 software includes the packet in a packet. The receiving device performs the same calcula­ frame for transmission to the destination DTE (user). At tion and compares its results with the value that the send­ the destination DTE (user), Level 2 software performs er­ ing device had placed in the FCS field. If these values do ror detection and sequence checking and strips accepted not agree, the frame has a transmission error and is dis­ frames down to the packet. Level 3 software is then applied carded. Discarding causes the receiving device to send an to the packet to strip the header information from the S-Frame with the Reject Recovery command. This packet and pass the user information to the appropriate S-Frame contains the frame numbered N(R), thus ac­ application program via the communications access knowledging receipt of all frames numbered N(R)-l and method. Table 7 summarizes the types of packets ex­ below, and initiates retransmission of frames N(R) and changed between terminals. above. The Packet Header consists of three octets. In Octet 1, , In effect, Level 2 envelops the packet in control infor­ the first four bits represent the Logical Channel Group ',,- ./ mation and prescribes procedures that ensure a high de­ gree of transmission accuracy and detection aflost frames during transmission.

AUGUST 1992 @ 1992 McGraw-Hili, Incorporated. Reproduction Prohibited. Datapro In!ormation Services Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 15 Networking Standard. Xlerie. Standards

Figure 5. ( X.25 Packet Header ForltUlt

·...... ----Octet 1 --~*I.----Octet 2--~*r----Octet 3----1

D-Bit Q-Bit M-BII

Bits 1 2 3 4:5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 Logical General Channel Format Logical Channel Number Packet Type Identifier Group Identifier Number

Number, and the final four bits represent the General For­ The destination DCE transmits an Incoming Call mat Identifier. Octet 2 represents the Logical Channel packet to the originating DTE and places the logical chan­ Number, and Octet 3 represents the Packet Type Identi­ nel in the DCE Waiting state. fier. See Figure 5. The called DTE indicates its willingness to accept the The Logical Channel Group Number and the Logical call by transmitting a Call Accepted packet across the Channel Number provide the routing information that di­ DTEIDCE interface. The Call Accepted packet specifies rects the packet over one of the logical channels. The num­ the same logical channel that is indicated by the Incoming bering system for the logical channels is dynamic and re­ Call packet. This places the specified logical channel in the fers to a switched data path within the packet network. Data Transfer state. (The logical channel at the calling The General Format Identi/ierindicates the general for­ DCE is still in the Wait state.) mat of the rest of the header. Specific bit patterns are es­ A Call Connected packet is then sent by the DCE to the tablished for the following: call setup packets; clearing, calling DTE and sets the logical channel status to the Data flow control, interrupt, reset, restart, and diagnostic pack­ Transfer state; the logical channel is then ready for transfer ets; and data packets. of data packets. This applies only for virtual circuit con­ The Packet Type Identifier establishes the packet's func­ nections; for permanent virtual circuit connections, the as­ tion. Examples of functions include call setup, priorities, signed logical channels are always in the Data Transfer and data. If applicable, it may identify the packet's place in state. sequence. See Table 7 for the various packet types. It is possible for a DCE to receive a Call Request (from a DTE) and an Incoming Call (from the network) simulta­ Packet-Level Procedures neously, with both messages specifying the same logical X.25 Level 3 defines procedures for call initiation, data channel. This is called a call collision; when a collision oc­ transfer, interrupts, reset, restart, and clearing. Some of curs, the DCE cancels the Incoming Call and proceeds to these procedures are summarized below. process the Call Request. Call Initiation: The Level 3 software in a DTE initiates Data Transfer. Once the logical channel is in the Data a transmission by sending a Call Request packet. This Transfer state, data, flow control, or reset information packet enables the calling DTE to request the opening of a packets can be transmitted. logical channel. The calling DTE designates the channel The Data packet format includes the packet send and number based upon the original assignments that were receive sequence numbers. The calling DTE can transmit made when the user subscribed to the network. The Call up to a predetermined number of packets before requiring Request packet also informs the network of the calling a response. This number is referred to as the window. All DTE's address and of the called DTE's address. Until the packet networks must accommodate a lower window edge call is disconnected, the network retains the addresses of of at least eight, permitting at least seven packets to be both devices associated with the logical channel number. outstanding before an acknowledgment is required. The Therefore, each data packet that is transmitted needs to upper window value (the maximum number of outstand­ contain only the logical channel number. The network ap­ ing packets) may not exceed 127 (modulo 128). pends the destination address just before routing the Upon receipt of an authorized packet, the receiving de­ packet. At the time the Call Request is issued, the logical vice can either authorize additional transmission by trans­ channel indicated by the calling DTE must be in a Ready mitting a Receive Ready packet to the sending device or state, that is, not being used to handle another call. Upon deny authorization to transmit by issuing a Receive Not ( receipt of the Call Request by the called DTE, the specified Ready packet. logical channel is designated as being in the DTE Waiting When transmitting Data packets, the Interrupt packet state. The packet is then transmitted to the destination can bypass flow control (authorization procedures). A DCE. DTE can transmit an Interrupt packet to another DTE. To avoid swamping the receiver with Interrupt packets, the

@ 1992 McGraw-HIII,lncorporated. Reproduction Prohibited. AUGUST 1992 Datapro Information Services Group. Delran NJ 08075 USA 18 2120 CCln Packet Switched Data Networking NetworIcInt SUnd.rd. Standards X ......

Figure 6. Basic System Structure ofX.7S SiglUlling IUId Data Tranafer Procedures r ------.... -- -- -I I Higher Level Functions Packet Packet I I I Link -Call Control Signaling ~ Transfer -Network Control Procedures Procedures I A1 or G1 -User Traffic I XIY Signaling Terminal (STE) I Interface

Gatewayffransit Data Switching Exchange X or Y

sender cannot send a second Interrupt packet until the first Implementation Considerations for Users Interrupt packet is acknowledged by an Interrupt Confir­ X.25 recognizes that DTEs differ in their degree of sophis­ mation packet. tication. A simple DTE may have fixed packet formats and While Data packets or Interrupt packets are being built-in parameter values, while a sophisticated DTE may transmitted, issuing a Reset Request packet can reset the work with varying packet formats and provide variable pa­ call. A DCE initiates a reset by issuing a Reset Indication rameters to take advantage of the many network functions packet. In either case, the logical channel is placed in the and signaling capabilities offered. Data Transfer state. Any Data, Interrupt, Receive Ready At the physical level, the transmission rate DTE uses to (RR), or Receive Not Ready (RNR) packets in the network access an X.25 network should be governed by the at the time of reset are ignored. If a DTE and DCE trans­ throughput and response time requirements of the user. mit Reset packets simultaneously, a reset collision occurs. The net effect is to place the logical channel in the Data Transfer state, ready for Data and Interrupt packets. Table 4. X.21 Pin Assignments A DTE or a DCE initiates a request for clearing (discon­ necting) a logical channel. Upon confirmation by the other Pin Employing Employing device, the logical channel is placed in the Ready state. Number X.26 X.27 Error Htmdling (Packet Level): When an error occurs, the DCE transmits Reset, Clear, and Restart indication 1 See Note (3) See Note (3) packets to the DTE. Reset reinitializes a virtual call or per­ 9 Ga T(B) manent virtual circuit. Reset removes all Data and Inter­ 2 T T(A) rupt packets on the logical channel and resets the lower window edge to zero. It affects only one logical channel 10 Ga C(B) number (one user). Reset procedures, which apply only in 3 C C(A) the Data Transfer state, handle specific error events, such 11 R(B) R(B) as local procedure error, remote procedure error, network congestion, incompatible destination, network out of or­ 4 R(A) R(A) der, etc. Clear affects only one logical channel number (one 12 I(B) I(B) user). It clears a session when, for example, the host or host 5 I(A) I(A) link crashes. In this case, the network initiates Clear proce­ 13 S(B) S(B) dures on the terminal link. It effectively logs off the termi­ nal user. Clear resets the lower window edge to O. Restart 6 S(A) S(A) affects all users (all logical channels on a given link) and 14 B(B) B(B) clears all calls on that link. Restart is used when a failed 7 B(A) B(A) link returns to service; it makes sure that all calls were cleared when the link went down. Restart also is used when 15 Reserved for Reserved for either side (DTE or DCE) reports error conditions at the future use future use packet level; the good side initiates the restart procedures. 8 G G In some networks, a diagnostic packet indicates unusual Notes: error conditions. A diagnostic packet from the DCE pro­ (1) X.21 pin assignments are defined by ISO 4903-1980. vides information on events that are considered unrecov­ (2) Where balanced clrults are, the associated pairs are deslg- erable at the packet level; the information permits the hated "A" and "B." DTE to analyze and initiate recovery by higher levels. No (3) Pin 1 Is reserved for connection of the shield or shielded In- confirmation is required, on receipt of a diagnostic packet. terconnecting cable.

AUGUST 1992 @ 1992 McGraw-HiII,lncorporated. Reproduction Prohibited. Datapro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 17 Net_riling Standard. X Serle. Standards

(- Table 5. X.25 Level 2 Commands and Responses

Commands (1) Responses Encoding of Control Field Format 1 2 3 4 5 6 7 '8 Information (information) 0 to- N(S)-f P f- N(R)-{ transfer Supervisory RR (3) (receive ready) RNR (receive not ready) 0 0 0 P/F N(R) RNR (3) (receive not ready) RNR (receive not ready) 0 1 0 P/F N(R) REJ (3) (reject) REJ (reject) 0 0 1 P/F N(R) Unnumbered SARM (set asynchronous OM (disconnected (2,3) response mode) (2) mode) P/F 0 0 0 SABM (set asynchronous (2) balanced mode) 1 1 P 1 0 0 DISC (disconnect) 0 0 P 0 0 UA (unnumbered acknowledgment) 0 0 F 1 1 0 CMDR (command reject) 0 F 0 0 FRMR (frame reject) (1) The need for, and use of, additional commands and responses are for further study. (2) DTEs do not have to implement both SARM and SABM; furthermore, DM and SABM need to be used if SARM only is used. (3) RR, RNR, and REJ supervisory command frames are not used by the DeE when SARM Is used (LAP).

Factors to consider include the maximum number of vir­ Recommendation X.75 tual circuits operating simultaneously and traffic charac­ Recommendation X.75, Terminal and Transmit Call Con­ teristics of throughput-critical and response-time-critical trol Procedures and Data Transfer System on International virtual circuits. Circuits Between Packet-Switched Data Networks, pro­ At the link level, Link Access Procedures (LAPs) and vides the rules for transmitting data between different data Link Access Procedures Balanced (LAPBs) are defined. A networks. The basic system structure is made up of com­ DTE might implement only LAPB. Parameters that are municating elements that function independently. These part of the LAPB procedures can be set to constants for all elements include the physical circuits, which comprise network connections. For example, a timer (TI) may be set links Al or G 1 (as defined in X.92), and a set of mechani­ according to the slowest required line speed; a constant cal, electrical, functional, and procedural interface charac­ such as three seconds may be used. Also, the maximum teristics; packet transfer procedures, which operate over permitted number of unacknowledged I-Frames (informa­ the physical circuits and provide for the transport of pack­ tion frames) may be determined. The network provider ets between STEs; and packet signaling procedures, which must agree to this. All networks support a window of at use the packet data transfer procedures and provide for the least seven I-Frames. The maximum packet size (number exchange of call control information and user traffic be­ of bits) of the I-Frame must also be established. tween STEs. Figure 6 shows the basic system structure of If the DTE handles certain character sizes (octets), the the signaling and data transfer procedures. The interna­ frame level should be capable of accommodating any num­ tionallink is assumed to be data link Al andlor data link ber of bits correctly (generate acknowledgment, calculate GI as defined in Recommendation X.92. The interna­ frame check sequence). How such a properly received tionallink should be capable of supporting full-duplex op­ frame is processed depends on the individual system and eration. may include actions such as discarding the frame, padding Link-level packet transfer procedures between STEs it, clearing the call, or resetting. consist of the Single Link Procedure (SLP) and the Multi­ link Procedure (MLP). The SLP is used for data inter­ Connections Between Packet Switched change over a single physical circuit between two STEs. When multiple physical circuits are used in parallel, the Data Networks SLP is used independently on each circuit. The MLP is CCITT Recommendation X. 7 5 describes the procedures used for data interchange over multiple parallel links. The for the interconnection of packet switched networks. MLP may be used over a single link when the communicat­ Many public packet switched data networks support X.75 ing parties agree to this procedure. Transmissions are full procedures, including AT&T Accunet, TransCanada's duplex. SLP is based upon the Link Access Procedure Bal­ Datapac, Graphnet Freedom Network II, US Sprint's anced (LAPB) as described in X.25 and uses the principle SprintNet, and WangPac networks. and terminology of the International Organization for X.75 is similar to X.25 in that it specifies procedures Standardization's (ISO's) High-level Data Link Control for the physical, link, and packet levels. Signal Terminal (HDLC). For SLP, either m.odulo 8 (nonextended mode) Equipment (STE), which acts as a bridge node between or the modulo 128 (extended mode) may be used. The ( networks, implements X. 7 5 procedures. A related stan­ MLP is based on the multilink procedures specified by the dard, CCITT X.121, defines the international numbering ISO and performs the functions of distributing packets plan for public data networks. across available SLPs. Three channel states are defined: the link channel state, which provides transmission in one direction; the active

@ 1992 MCGraw-Hili, Incorporated. Reproduc1lon Prohibited. AUGUST 1992 Datapro Information Services Group. Delran NJ 08075 USA 18 2720 eelTT Pack.' Switched Data Networking Networking Standarda Standards XS.rI••

Table 6. Summary of Packets Exchanged aetween Terminals During a Virtual Call

Events Activity at Calling DTE Activity at Called DTE Can Initiation Can Request packet is sent Incoming Can packet is received Can Accepted packet is sent Can Connected packet is received

Data Transfer Data packet sent Data packet received Ready Receive packet sent Ready Receive packet received Data packet B sent" Data packet A sent" Data packet A received" Data packet B received"

Disconnect Clear Request packet sent Clear Indication packet received Clear Confirmation packet sent Clear Confirmation packet received "Two-way (full-duplex) transmission of data packets between terminal equipment.

channel state, which means that the incoming or outgoing STE by setting the P bit to 1; the answering STE responds channel is receiving or transmitting a frame; and the idle by setting the F bit to 1. The following commands and re­ channel state, which means that the incoming or outgoing sponses are supported: channel is receiving or transmitting at least 15 contiguous • Information (I) command-transfers sequentially num­ 1 bits. bered frames that contain an information field; Data is transmitted in frames. Each frame must contain an Opening Flag (8 bits), an Address field (8 bits), a Con­ • Receive ready (RR) command and response-used by the trol field (8 bits), a Frame Check Sequence (FCS) field (16 STE to indicate that it is ready to receive an I frame or to bits), and a Closing Flag (8 bits). An Information field of acknowledge a previously received I frame; an unspecified number of bits, which follows the Control • Receive not ready (RNR) command and response-used by field, is optional. the STE to indicate a busy condition; The Control field contains a command or response, and sequence numbers if applicable. Control field formats may • Reject (REJ) command and response-used by the STE to be one of three types: numbered information transfer (I request retransmission of I frames beginning with frame format), numbered supervisory functions (S format), and numbered N(R); unnumbered control functions (U format). The I format • Set asynchronous balanced mode (SABM) command-an performs information transfer functions. The S format unnumbered command used to set the addressed STE in performs link. supervisory control functions, such as ac­ the asynchronous balanced mode information transfer knowledging I frames, requesting transmission of I frames, phase; all command/response control fields are one octet and requesting a temporary suspension of transmission of (eight bits); I frames. The U format provides additional link control • Set asynchronous balanced mode extended (SABME) com­ functions. mand-an unnumbered command used to set the ad­ Each I frame is numbered sequentially. The send state dressed STE in the asynchronous balanced mode infor­ variable V(S) represents the sequence number of the next mation transfer phase; all numbered command/response in-sequence I frame to be transmitted. The value of the control fields are two octets; all unnumbered command/ V(S) is incremented by one with each successive I frame response control fields are one octet; transmission but cannot exceed N(R) of the last received I or S format frame by more than the maximum number of • Discon,nect (DISC) command-terminates the previ­ outstanding frames. The send sequence number N(S), con­ ously set mode; DISC indicates that the STE transmit­ tained only in 1 frames, is set equal to the value of V(S). ting DISC is suspending operation; The receive state variable V(R) represents the sequence • Unnumbered acknowledge (UA) response-used by the number of the next in-sequence I frame expected to be re­ STE to acknowledge mode-setting commands; ceived. The value of V(R) is incremented by one when an error-free, in-sequence I frame whose N(S) equals V(R) is • Disconnected mode (DM) response-reports STE status received. Both I frames and S frames contain the receive when it is logically disconnected from the link and is in sequence number N(R). N(R) is the expected send se­ the disconnected phase; and quence number of the next received I frame. When the • Frame reject (FRMR) response-indicates an error condi­ STE transmits N(R), it indicates that all I frames num­ tion that is not recoverable by retransmission of the bered up to and including N(R)-1 have been received cor­ frame. rectly. All frames contain the Poll/Final (P/F) bit; which is re­ Packet-level signaling procedures relate to the transfer of ferred to as the P bit in command frames and the F bit in packets at the STE-X/STE-Y (XIY) interface. Recommen­ response frames. The STE solicits a response from another dation X. 7 5 specifies signaling procedures for virtual call

AUGUST 1992 @ 1992 McGraw-Hili, Incorporated. Reproduction Prohibited. Datapro Information Servloes Group. Delran NJ 08075 USA Data Networking CCITT P.cket Switched 2720 19 Networking Stend.rd. XSerie. Standards

( Table 7. Packet Types and Their Use in Various Functions

Packet Type Service

Function From DCE to DTE From DTE to DCE Switched Perm. Virtual Virtual Circuit Circuit Call Setup and Cleaning In Incoming Can Can Request X can Connected Call Accepted X Clear Indication Clear Request X DCE Clear DTE Clear Confirmation Confirmation X Data and Interrupt DCE Data DTE Data X X DCE Interrupt DTE Interrupt X X DCE Interrupt DTE Interrupt Confirmation Confirmation X X Flow Control and Reset DCE RR (Receive DTE RR X X Ready) DTE RNR X X DCE RNR (Receive DTE REJ* X X Not Ready) Reset Request X X Reset Indication DTE Reset DCE Reset Confirmation X X Confirmation Restart Restart Indication Restart Request X X DCE Restart DTE Restart Confirmation Confirmation X X Diagnostics Diagnostics X X

setup and clearing, for permanent virtual circuit service, state exists within the packet-level ready state of a logical for data and interrupt transfer, for flow control, and for channel. Each data packet contains a sequence number reset. called the packet send sequence number P(S); only data Logical channels are used to complete simultaneous vir­ packets contain the P(S). tual calls and/or permanent virtual circuits. A logical chan­ The procedures for flow control and reset apply only in nel group number and a logical channel number (in the the data transfer state. Flow control applies to data pack­ range 0 to 15 inclusive and 0 to 255 inclusive, respectively) ets. A window is defined for each direction of transmission are assigned to each virtual call and permanent virtual cir­ at the XlY interface. The lowest number in the window is cuit. The logical channel group number and the logical called the lower window edge. The maximum window edge channel number are contained in each packet type, except does not exceed modulo 8 or 128, is unique to each logical restart packets. channel, and is reserved for a period of time. For a partic­ The procedures for virtual call setup and clearing are ular call, two window sizes, one for each direction of trans­ used only when a logical channel is in the packet-level mission, may be selected. The packet receive sequence ready (RL) state. If call setup is possible, and no call or call number P(R) (modulo 8 or 128) conveys information from attempt exists, the logical channel is in the ready (PL) state the receiver for the transmission of data packets. The P(R) (within the RL state). Call setup is initiated when the STE becomes the lower window edge when transmitted across sends a call request packet across the XIV interface. The the X/y interface, thereby authorizing additional data call request packet specifies a logical channel in the PL packets to cross the X/Y interface. state. The logical channel is then placed in the call request Reset procedures are used to reinitialize a single call state. The called STE indicates acceptance by the called and apply only in the data transfer state. The STE sends a DTE by sending a call connected packet across the X/y reset request packet that specifies the logical channel to in­ interface. It specifies the same logical channel as that re­ dicate a request for reset. The logical channel is placed in quested by the call request packet. The logical channel is the reset request state. The requested STE confirms by then placed in the flow control ready (DL) state within the sending a reset confirmation packet, which places the logi­ data transfer (P4) state. cal channel in the flow control ready state. A logical channel (in any state) can be cleared when the Restart procedures are used to clear all calls simulta­ STE sends a clear request packet, which specifies the logi­ neously. When the STE sends a restart request packet, the cal channel, across the XlY interface. Upon receipt of a X/Y interface for each logical channel is placed in the re­ clear request packet, STE-X or STE-Y frees the logical start request state. In this state all packets, except restart channel and transmits a clear confirmation packet that request and restart confirmation packets, are discarded by specifies the same channel. This places the logical channel the X/Y interface. An STE confirms by sending a restart in the ready state within the RL state. Permanent virtual confirmation packet, which places all channels in the ready circuits require no call setup or clearing. state. Data transfer procedures apply independently to each Packet formats are based on the general structure of logical channel at the XlY interface. In the data transfer packets as defined in X.25. (P4) state, data, interrupt; flow control, and reset packets may be sent and received by the STE. The data transfer

@ 1992 McGraw-Hili. Incorporated. Reproduction Prohibited. AUGUST 1992 Datapro InfOrmation Services Group. Delran NJ 08075 USA 20 2720 cCln Packet SWItched Data Networking Networking Standard. Standards XSerie.

Trends in Packet Switching prior to exchanging data. In a connectionless service, typi­ cal of local area networks, each packet is independently X.25 packet switching is widely supported in existing data routed to the destination. No connection establishment ac­ processing and data communications equipment. All ma­ tivities are required since each data unit is independent of jor host computer and communications processor ven­ the previous or subsequent one. Each transmission mode dors, for example, have incorporated X.25 interfaces into has a niche where it represents the best approach. For ex­ their products. This is part of an overall trend in accepting ample, file transfers may benefit from a connection-ori­ international standards and the increasing availability of ented service, while point-of-sale inquiries may be best products conforming to these standards. served by a connectionless service. The CCITT published revisions to the X Series stan­ Traditionally, the CCITT has pursued a connection­ dards in 1984 and in 1989. Since that time, the ratification oriented philosophy, while ISO has shown interest in con­ and publication of revisions has become a continuous, on­ nectionless. While the original OSI standard, ISO 7498, is going process. Since the major building blocks for X.25 connection oriented, ISO saw the need to provide connec­ were laid by 1984, all subsequent changes have been, and tionless serVice by issuing an addendum to that protocol, will continue to be, relatively minor. Some post-1984 ISO 7498/DADl. ISO has issued a standard for connec­ changes have revolved around efforts to make CCITT X tion-mode network service (ISO 8348), while the CCITT Series standards compatible with those of the Interna­ has issued an identical service, X.2l3. In regard to X.25 tional Organization for Standardization (ISO). Currently, itself, however, ISO has decided not to pursue the connec­ discussions on how to provide greater interoperability be­ tionless service, formerly known as "datagram" service. tween various X.25 networks is taking place. Major devel­ X. 7 5 is also a connection-oriented service; ISO has shown opments in packet switching today, however, center not considerable interest in a connectionless internetworking around X.25, but around the development of new ISDN­ protocol (IP) and has developed the ISO 8473 to accom­ related technologies, such as frame relay and Broadband modate it. ISDN, which provide much higher throughput through simplified packetization and routing schemes. This section Packet Switching in ISDN discusses post-1984 changes to X.25 and its relationship to The goal of the Integrated Services Digital Network ISDN. (ISDN) is to provide an end-ta-end digital path over a set of standardized user interfaces, giving the user the capabil­ Major Changes in 1988 Revisions ity to signal the network through an out-of-band channel. In the 1988 revised standards, there were no changes at the (In contrast, in X.25, the user signals the network in in­ physical and link levels. At the packet level, however, a band fashion by issuing packets such as CALL REQUEST, new facility for redirecting calls, Call Deflection, was estab­ CALL ACCEPTED, etc.) lished. In 1984, the CCITT had made available a new Call Currently, different types of interfaces to the telephone Redirection facility, allowing the network to redirect all network exist for different services. These interfaces in­ calls destined for a given address. This redirection could clude two-wire switched, twa-wire dedicated, four-wire occur when the destination was out of order or busy, or it dedicated, DDS, and so forth. ISDN will provide a small could be based on time of day or other criteria. The 1988 set of interfaces that can be used for multiple applications. facility extended this capability, allowing the destination The CCITT has defined the following interfaces for ISDN: subscriber to clear incoming calls to another party on a call-selective basis. The Clear Request packet contains the • 2B+D...,-two 64K bps channels and a 16K bps packet/ Call Deflection information that profiles the desired alter­ signaling channel (also called the Basic Rate Interface). nate party. • 23B+D-twenty-three 64K bps channels and a 64K bps packet/signaling channel (also called the Primary Rate Relationship Between CCITT and ISO Efforts Interface). CCITT X.25 packet-level protocol specifies a virtual cir­ • 3HO+D-three 384K bps channels and one 64K bps cuit service; the ISO has issued a compatible version of the packet/signaling channel. packet standard, ISO 8208. In recent years, CCITT and ISO organizations have worked on standards to carry • H11-nonchannelized 1.536M bps. longer addresses in the DTE field to facilitate interworking • H12-nonchannelized 1.920M bps. with ISDN (E.164). In 1988, the CCITT also modified the Address Exten­ • Multislotted-multiples of 64K bps channels (up to sion facilities to be consistent with ISO address length. 1.536M bps) under the customer's control. Previously, a provisional 32-decimal/16-octet field had • Broadband-high data rates, based on an approach been recommended; this address length was increased to called synchronous optical network (SONET), building 40 decimalsl20 octets. The ISO also added these address on multiplex of 51.84M bps. SONET standards negotia­ recommendations to Addendum 2 of ISO 8348 (Connec­ tions began in 1986. The CCITT approved phase I of the tion-Mode Network Service); the CCITT adoption is standards in 1988 and phase II in 1989. This architecture X.213. has been called Broadband ISDN, in contrast to the other interfaces that have been considered part of Nar­ Connectionless Issues: Relationships With ISO Efforts rowband ISDN. At any layer of the Open Systems Interconnection (OSI) Reference Model, two basic forms of operation are possi­ With the exception of Broadband ISDN, all of the above ble: connection oriented and connectionless. Connection­ interfaces could be carried on unloaded copper loops. Us­ oriented service involves a connection establishment ing fiber has also been considered, as it would make the phase, a data transfer phase, and a connection termination phase. A logical connection is set up between end entities

AUGUST 1992 @ 1992 McGraw-Hili, Incorporated. Reproduction Prohibited. Datapro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Packet SwHched 2720 21 Networking Standard. XSerie. Standards

local loop more robust. Out-of-band signaling makes pos­ physical interface. Second, each user device may support sible a new class of services. In addition, the 16K bps multiple types of traffic, including packet switched data ( D-channel will be connected directly to the BOC's packet and signaling. To accomplish this type of multiplexing, switched network, providing the subscriber with the data LAP-D employs a two-part address consisting of a termi­ multiplexing advantages packet switching offers. A major nal endpoint identifier (TEl) and a SAP!. Typically, each effort is under way in Europe to bring the system to mar­ user terminal is given a distinguishing TEL The SAPI iden­ ket. In the United States, several trials have been under­ tifies the traffic type and the Data Link Layer services di­ taken, and limited ISDN service is already available. rected to Layer 3. For example, the SAPI value of 0 directs the frames to Layer 3 for call-control procedures; a SAPI CCnT ISDN Standards value of 16 indicates a packet communication procedure. ISDN provides a specific protocol that users can employ to See Figure 7. signal the network. Currently, a three-layer protocol suite is defined. At the Basic Rate, the Physical Layer manages a Frame Relay 192K bps, full-duplex bit stream using time-compression Frame relay is a rapidly emerging, standards-based ad­ techniques and time division methods to recover the two dressing technique that has great potential in LAN/WAN B-channels and one D-channel. The remaining 48K bps networking and other interactive applications requiring stream is used for Physical Layer control information. The high-throughput, low-delay transmission. Frame relay is defined standards are 1.420 (Basic Rate Interface Defini­ based on the CCITT Layer 2 protocol developed for ISDN, tion), 1.430 (Basic Rate Interface Layer 1), 1.421 (Primary Link Access Protocol D (LAPD). Unlike conventional Rate Interface Definition), and 1.431 (Primary Rate Inter­ X.25 packet switching, frame relay uses variable packet face Layer 1). lengths and performs error checking only at the remote end The Data Link Layer is not defined for transparent of transmission. Any errors occurring between intermedi­ B-channels used for circuit switched voice or data, but it is ate network nodes are assumed caught and corrected by defined for the D-channel. For Narrowband ISDN, the higher-layer protocols. Thus, intermediate nodes simply D-channel employs a LAP-D Link Layer protocol, which is forward packets (called frames) without processing the a subset of the ISO HDLC Data Link protocol, as specified datastream. In addition, frames must be received in the in CCITT Recommendations Q.920 (1.440) and X.921 order in which they were sent, unlike some X.25 networks, (1.441). It provides statistical mUltiplexing for three chan­ which involves considerably less machine processing at the nel types: signaling information for the management ofthe opposite end of transmission. These efficiencies result in B-channels; packet switched service over the D-channel; superior performance, with data rates up to TlIEllevels. and optional channels, used for telemetry of other applica­ Vendors from several different networking disciplines tions. have established themselves as frame relay equipment pro­ The Network Layer protocol for the signaling channel is viders. These include traditional packet switched equip­ specified in CCITT's Q.930 (1.450) and Q.931 (1.451) ment suppliers such as Northern Telecom, US Sprint, specifications. It provides the mechanism for establishing BBN Communications, BT North America, Hughes Net­ and terminating connections on the B-channels and other work Systems, and Netrix Corp.; T-carrier nodal processor network control functions. For the packet switched service vendors such as General DataComm, Newbridge Net­ over the D-channel, the Network Layer protocol is X.25. works, Network Equipment Technologies, StrataCom, and CCITT will define Layer 3 protocols for the optional chan­ Timeplex; LAN internetworking (bridge and router) ven­ nels in the future, or they will be specified as national op­ dors; and communications carriers. tions. Most commercial frame relay products are based on A technique is required for specifying whether user-to­ ISDN recommendations contained in the following Amer­ network signaling, user packet data, or user telemetry data ican National Standards Institute (ANSI) specifications: is being sent over the D-channel. This technique involves • TI.606: Frame Relaying Bearer Service-Architectural the use of a service access point identifier (SAPI). Framework and Service Description (TlS1I88-225) Each layer in the OSI Reference Model communicates with the layers above and below it across an interface. The • Addendum to TI.606: Frame Relaying Bearer Service­ interface is through one or more service access points Architectural Framework and Service Description (SAPs). SAPs have a number of uses, including subad­ (TlS1I90-175) dressing for internetworking situations, Transport Layer • TI.6ca: Core Aspects of Frame Protocol for Use with applications, and for user data packet service over an Frame Relay Bearer Service (TlS1I90-214) ISDN D-channel. Considering the applications to the Transport Layer, The CCIn has released frame-relay specification 1.122, en­ one should note that two general types of addressing in a titled "Framework for Providing Additional Packet Mode communications architecture are available. Each host on Bearer Services." Currently, only one packet service has the network must have a network address, allowing the been specified in ISDN standards: "Support of Packet network to deliver data to the proper computer. Each pro­ Mode Terminal Equipment by an ISDN," CCITT 1.462 cess within a host must have an address that is unique (X 31). within the host; this allows the Transport Layer to deliver data to the proper process. These process addresses are Broadband ISDN and Cell Relay identified using SAPs. A similar approach is followed for Broadband ISDN (BISDN) is the blueprint for public net­ ISDN. works in the mid-1990s (1994 and beyond). It is being de­ LAP-D, the data link standard for ISDN, specifies the veloped to support switched (on demand), semiperma­ ( link access protocol used on the D-channel. LAP-D is nent, and permanent broadband connections for both based on LAP-B, which is based on HDLe. LAP-D must point-to-point and point-to-multipoint applications. deal with two levels of multiplexing. First, at a subscriber Channels operating at 155M bps and 622M bps will be location, multiple-user devices may be sharing the same

@ 1992 McGraW-Hili, Incorporated. Reproduction Prohibited. AUGUST 1992 Datapro Information Services Group. Delran NJ 08075 USA 22 2720 ccm Packet Switched Data Networking Networking Standard. Standards X Serle.

Figure 7. SAP] Action for a Basic Rate D-Channel

Signaling Packet Telemetry ------Layer 4

Layer 3 Layer 3 Network Layer 3 Layer 3 Layer s Entity p Entity 3 Layer 81 Entity 82 Entity t Entity (SAP1=O) (SAP1=16) ~---Pro~COJ----~

Layer 3 ------.------.------Layer 2 Layer 2 Layer 2 Data Link 81 Entity 82 Entity LAPD Entity Layer 64K bps 64Kbps 64K bps ~----prolooo~----~

_____ 1 Layer 2 , ------;------~------. Layer 1 ISDN Physical Entity Physical Layer Protocols 192K bps ~------~

Layer 1

Physical Media (ISDN Link) ~------~ available under BISON, allowing transmission of data, by 512 150M bps lines. Major trials are planned for 1992 video, and digitized voice. and beyond; a trial scheduled in Belgium in 1992 follows Broadband services are aimed at both business applica­ 1991 trials in the U.S. and Japan. Several high-speed tions and residential subscribers. Connections will support switching trials in the past two-to-three years have been both circuit mode and packet mode services of a single undertaken in the U.S. and in Germany. NYNEX is plan­ media, mixed-media, and multimedia. BISON's founda­ ning a trial in Boston while BellSouth plans one in North tion is cell relay, and particularly the international stan­ Carolina. MCI Communications Corp. plans an early de­ dard supporting it: Asynchronous Transfer Mode (ATM). ployment of Siemens Stromberg-Carlson ATM switches to Cell relay is a high-bandwidth, low-delay switching and provide broadband interLATA services in the U.S. These multiplexing technology in which information is pack­ services are expected for the 1992-93 time frame. In late etized into fixed-size slots called cells. A cell consists of an 1991, Southern Bell Telephone installed what was hailed information field that is transported transparently by the as the "first broadband ISDN switch in a U.S. CO." The network and a header containing routing information. switch, which supports the 622M bps UNI, will be used by With simplified protocols and cells with a fixed, short the University of North Carolina, as part of the VISTA­ length (53 bytes), cell relay will make very high data rates NET undertaking. possible. A number of cell switch vendors have efforts underway The CCITT has issued several Broadband ISDN speci­ to develop ATM-based equipment for frame relay applica­ fications in its I-Series recommendations. 1.361, the tions, rather than for mixed-media CO applications. It ap­ BISDN ATM Layer Specification, defines the ATM cell pears that, in the U.S., the route to frame relay will be via structure and coding, including header formats and coding cell relay switches; these support frame relay interfaces on at both the User-Network Interface (UNI) and Network the access side and cell relay on the backbone side. Exam­ Node Interface (NNI). It also defines ATM protocol proce­ ples of this type of equipment include AT &Ts BNS-l 000 dures. I. 311, entitled BISDN General Network Aspects, de­ and BNS-2000, and Stratacom IPX products. StrataCom scribes networking techniques, signaling principles, traffic supports a 24-octet cell (with 4 octets of overhead); AT&T control, and resource management for BISDN. It defines supports a 16-octet cell on the BNS-l000. Other early ATM virtual section, virtual path, and virtual channel con­ ATM equipment or service entrants include: Fujitsu Net­ cepts. work Switching, ASCON Timeplex (TIMEPATHlEsprit), Off-the-shelf ATM products for BISDN are expected by Network Equipment Technologies, Ungermann-Bass Inc., 1995; some ATM products are available already and a BBN (Emerald), and Siemens Stomberg-Carlson (Metro­ great deal of work continues world-wide. In 1990, Fujitsu politan Area Network Switching System). Several of these announced a commercial BISON switch. It switches 512 switches support both voice and data.•

AUGUST 1992 @ 1992 McGraw·HIII, Incorporated. Reproduction Prohibited. Datapro Information Services Group. Delran NJ 08075 USA DATAPRO Data Networking 2720 1 Standards

~- CCITT Packet Switched Networking Standards X Series

In this report: Synopsis

Packet AssemblYI Editor's Note Report Highlights Disassembly...... 3 In 1984, CCITT published Red A packet switched network permits a Books on wide-ranging topics, in­ user's data terminal equipment (i.e., X.21 Interface cluding the X.25 packet-switching a PC, host computer, or terminal) to Specifications ...... ·7 standards. A set of revisions to the X communicate with the equipment of Series, the Blue Books, was pub­ other geographically dispersed users. Recommendation lished in 1989. Since that time, stan­ Data must be presented to the net­ X.25 ...... 13 dards ratification and publication work in a prescribed manner, how­ have been ongoing, continuous pro­ ever. A packet assembler/ Connections between cesses. Each future addition or revi­ disassembler (PAD), also referred to Packet Switched Data sion to the X Series standards will be as data circuit-terminating equip­ Networks ...... ·23 made available, as soon as they are ment (DTE), serves as a network finalized, in individual grey booklets. entry/exit point, packetizing and de­ Trends in Packet Since the major building blocks of packetizing data according to the Switching ...... 25 the X standards were completed by rules specified by the X.3, X.28, 1984, all post-1984 technical changes X.29, X.21, X.25, and X.75 recom­ are relatively minor; they are dis­ mendations of the CCITT. cussed in this report, however.

Major developments in packet switching center around the develop­ ment of ISDN-related technologies, such as fast packet switching and frame relay, which provide integra­ tion of voice, video, and data, and support much higher throughput than traditional X.25 networks. IS­ DN's relationship with traditional X.25 packet switching is also dis­ cussed.

-By Martin Dintzis Assistant Editor

@ 1991 McGraw·HiII, Incorporated. Reproduction Prohibited. FEBRUARY 1991 Datapro Information Services Group. Calran NJ 08075 USA 2 2720 ccln Packat .wftc..... Data Networking Networking Standa_ Standards X ..rI ..

devices on the network through a physical connec­ tion to the PDN. Each of these protocols used a \.~ multiplexing technique that enabled a computer to Analysis establish and maintain one or more virtual circuits to other network communicating equipment. No industry standard for packet switching existed, however, and most computer manufacturers were reluctant to provide the necessary software to han.. In the early days of packet switching, each Public dIe the variety of network access protocols. Data Network (PDN) defined its own network ac­ With the adoption of the X Series Recom­ cess protocol, which permitted an appropriately mendations by the CCITT in 1976, the PDNs equipped computer to communicate with other could offer a standard network access protocol. The recommendations are continually fine-tuned.

Table 1. CCIT Recommendations-X Series

CCITT Description CCITT Description Recommendation Recommendation X.1 International user classes of service in X.27 Electrical characteristics for balanced public data networks: class 8 (2400 double-current interchange circuits for bps); class 9 (4800 bps); class 10 data communications equipment (9600 bps); or class 11 (48.000 bps) X.28 DTE/DCE interface for asynchronous X.2 International user facilities in public device access to the PAD facility of a data networks public data network in the same X.3 Packet assembly/disassembly (PAD) fa- country cUity in a public data network; lists op- X.29 Procedure for the exchange of control tions and defaults for interactive information and user data between a asynchronous terminal connection to packet mode DTE and a PAD facility X.25 packet networks X.32 Procedure for communications between X.4 General structure of signals of Interna- users and packet networks through tional Alphabet No.5 (IA5) code for the switched telephone network and data transmission over public data through circuit switched public data networks (IA5 is described in CCITT networks V.3) X.75 Expanded X.25 recommendation for in- X.20 Interface between data terminal equip- ternetwork communications between ment (DTE) and data clrcuit-termlnat- packet switched networks; interface Is ing equipment (DCE) for async defined between two STEs (Signaling transmission services on public data Terminal Equipments) that are a part networks of ISDEs (International Data Switching X.20 bis V.21-compatible interface between DTE Exchanges). with expanded support and DCE for async transmission ser- for wideband links. extended sequenc- vices on public data network ing. and an expanded network utility field for international call X.21 General-purpose interface between establishment DTE and DCE for synchronous opera- tion on public data networks X.92 Hypothetical reference connections for public synchronous data networks X.21 bis For use on public data networks by DTE that are designed to interface to X.95 Network parameters in public data synchronous V-Series modems networks X.24 List of definitions of interchange circuits X.96 Call progress signals in public data between DTE and DCE on public data networks networks X.121 International numbering scheme for X.25 Interface between DTE and DeE for ter- multinetwork communications contain- minals operating in the packet mode ing .. 4-digit DNIX (Data Network Iden- on public data networks tification Code). a-digit area code. 5- digit host identification. a 0- to 2-digit X.26 Electrical characteristics for unbalanced subaddress ',,- double-current interchange circuits for data communications equipment

FeBRUARY 1991 @ 1991 McGraW-Hili, Incorporated. F!eprQduction Prohibited. Datapro Information Services Group. Dalran NJ 08075 USA Data Networking CCITT Packet Switched 2720 3 Networking Standard. X Serle. Standards

Revised editions are published at four-year inter­ asynchronous terminal's characteristics. The vals; the 1989 draft incorporates the latest revi­ PAD's basic functions include: sions and recommendations. The next published • The assembly of characters into packets; revision will be available in 1992. This report focuses on Recommendations • The disassembly of the user data field; X.3, X.28, and X.29 (informally called the Interac­ • Virtual call setup, clearing, resetting, and inter­ tive Terminal Interface [ITI] standards); X.21; rupt procedures; X.25; and X.75. • Generation of service signals; • A mechanism for forwarding packets when the proper conditions exist; Packet Assembly/Disassembly Recommendations X.3, X.28, and X.29 define the • A mechanism for transmitting data characters, procedures by which asynchronous terminals, com­ including start, stop, and parity elements; puters, and other devices, often referred to as data • A mechanism for handling a break signal from terminal equipment (DTE), communicate with an asynchronous terminal; other devices via a packet switched network. • Editing of PAD command signals; Packet assemblers/disassemblers, also referred to • A mechanism for setting and reading the cur­ as DTE, commonly serve as network entry/exit rent value of PAD parameters; points. X.3 defines the basic and user-selectable func­ • A mechanism for the selection of a standard tions ofa packet assembler/disassembler (PAD). It profile (optional); also lists 22 parameters necessary to characterize a • Automatic detection of data rate, code, parity, specific device (e.g., bit rate, the escape character, and operational characteristics (optional); and and flow control technique). The proper setting of • A mechanism for the remote DTE to request a these values enables the PAD to correctly interpret virtual call between an asynchronous terminal the communicating device and vice versa. and another DTE (optional). X.28, a related standard, defines the proce­ dures for character interchange and service initial­ The PAD's operation depends on the selectable ization, the exchange of control information, and values of internal variables called PAD parameters. the exchange of user data between an asynchronous A set of parameters exists independently for each terminal device and a PAD. X.29 defines the pro­ asynchronous terminal. The current value (the bi­ cedures for the exchange of PAD control informa­ nary representation of the decimal value) of each tion and the manner in which user data is PAD parameter delimits the operational character­ transferred between a packet mode DTE and a istics of the related function. The initial value of PAD or between two PADS. each parameter is set according to a predetermined set of values, the initial standard profile. Twenty­ Recommendation X.3 two PAD parameters have been standardized by CCITT Recommendation X.3, Packet Assembly/ the CCITT. They are as follows: Disassembly Facility in a Public Data Network, out­ lines the procedures for packet assembly/ • PAD recall using a character-allows an asyn­ disassembly in asynchronous transmissions. These chronous terminal to initiate an escape from the functions can be programmed and built into a data transfer state or the connection-in-progress microprocessor-based "black box" that is placed state in order to send PAD command signals. between the terminal and the X.25 network at ei­ This parameter has the following selectable val­ ther the customer's premises or the entry point of ues: not possible, possible by character 1/0 the network node. (DLE), or possible by a user-defined graphics The PAD performs a number offunctions, character. some of which allow it to be configured, by either • Echo-enables characters received from the ( an asynchronous terminal device or another (re­ asynchronous terminal to be interpreted by the mote) PAD, so that its operation is adapted to the

@ 1991 McGraw-Hili, IncOrporated. Reproduction Prohibited. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA 4 2720 CCITT Packet Switched Data Networking Networking Standard. Standards X Se,le.

PAD and transmitted back to the asynchronous • Padding after carriage return-permits the PAD terminal. Selectable values are no echo (0) and to automatically insert padding characters in echo (1). the character stream sent to the asynchronous • Selection ofdata forwarding characters-allows terminal after the occurrence of a carriage re­ the asynchronous terminal to send defined sets turn character. This enables the asynchronous of characters. which the PAD recognizes as an terminal printing device to perform the carriage indication to complete the packet assembly and return function correctly. A value between 0 to forward a complete packet sequence as de­ and 255 indicates the number of padding char­ fined in X.25. The basic functions of this pa­ acters the PAD will generate. rameter are encoded and represented by a • Line folding-perm its the PAD to automatically decimal value. The functions include no data insert appropriate format effectors in the char­ forwarding character (represented by decimal acter stream sent to the asynchronous terminal. 0); alphanumeric characters A-Z, a-z. and 0-9 No line folding or a predetermined maximum (decimall); CR (decimal 2); ESC, BEL, ENQ. number of graphics characters per line may be and ACK (decimal 4); DEL, CAN, and DC2 selected. (decimal 8); ETX and EOT (decimal 16); HT. • Binary speed-a read-only parameter that nei­ LF, VT, and FF (decimal 32); and all other ther DTE can change. It enables the packet characters in columns 0 and I of International mode DTE to access a characteristic (known by Alphabet No.5 (IA5) not included in the above the PAD) of the asynchronous terminal device. (decimal 64). Speeds from 50 bps to 64K bps are represented. • Selection ofidle timer delay-permits the selec­ • Flow control of the PAD by the start/stop mode tion of the duration of a time interval between DTE-governs flow control between the asyn­ successive characters. When data received from chronous terminal and the PAD. The asynchro­ the asynchronous terminal exceeds this interval, nous terminal transmits special characters to the PAD terminates the assembly of a packet indicate whether it is ready to accept characters and forwards it as defined in the X.25 protocol. from the PAD. In IA5, these special characters • Ancillary control-defines flow control between switch an ancillary transmit device on and off. the PAD and the asynchronous terminal. Deci­ Decimal 0 represents no use of X-on (DC I) and malO represents no use of X-on (DC I) and X­ X-off (DC3); decimal 1 represents use of X-on I off (DC3); decimal 1 represents use of X-on/X­ X-off. off (data transfer); and decimal 2 represents the • Line-feed insertion after carriage return­ use of X-on IX-off (data transfer and command). permits the PAD to automatically insert a line­ • Control ofPAD service signals-provides the feed character in the character stream sent to or asynchronous terminal with the capability to received from the asynchronous terminal or af­ decide whether and in what format PAD service ter echo of each carriage return character. This signals are transmitted. function applies only in the data transfer state. • Selection ofoperation ofthe PAD on receipt of • Line-feed padding-permits the PAD to auto­ the break signal-after receiving a break signal matically insert padding characters in the char­ from the asynchronous terminal, the PAD may acter stream transmitted to the asynchronous do nothing, send an interrupt packet to a packet terminal after the occurrence of a line-feed char­ mode DTE or another PAD, reset, or send an acter. This enables the asynchronous terminal indication ofbreak PAD message to a packet printing mechanism to perform the line-feed mode DTE or another PAD. operation correctly. This function applies only • Discard output-permits a PAD to discard the in the data transfer state. content of user sequences in packets rather than • Editing-enables character delete, line delete, disassembling and transmitting them to the and line display editing capabilities. During the asynchronous terminal. Selections include nor­ PAD command state, the editing function is al­ mal data delivery or discard output. ways available; use or nonuse of the editing function in the data transfer state is selectable.

FEBRUARY 1991 @ 1991 McGraW-Hili, Incorporated. Reproduction Prohibited. Datapro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched Networking Standard. 2720 5 X Serle. Standards

• Character delete, line delete, and line display­ • Procedures for the exchange of user data be­ ( all editing functions represented by one user­ tween an asynchronous terminal and a PAD. selectable character from IA5. • Editing PAD service signals-enable the asyn­ Modems standardized for use on public switched chronous terminal to edit PAD service signals or leased line facilities establish the procedures for for printing devices and display terminals; also providing an access path (DTE/DCE interface). used for editing via one character from IA5. Procedures for both V and X Series interfaces are Editing is not selectable. defined. Transmission speeds up to 1200 bps are spec­ • Echo mask-when echo is enabled, echo mask ified for V-Series interfaces; they are in accordance designates that selected defined groups of char­ with either the V.21, V.22, or V.23 standard, de­ acters sent by the asynchronous terminal are not pending on facility type and speed. The V-Series transmitted back. The following may be se­ specifications define the procedures for setting up lected: no echo mask; no echo of CR; no echo of and disconnecting the access information path by LF; no echo ofVT, HT, and FF; no echo of BEL both the DTE and the PAD. and BS; no echo of ESC and ENQ; no echo of X-Series interfaces also are used with ACK, NAK, STX, SOH, EOT, ETB, and ETX; switched or leased line facilities. The physical char­ no echo of editing characters; or no echo of all acteristics for the DTE/DCE interface are specified other characters. in X.20 or X_20 bis. Procedures for setting up and • Parity treatment-permits the PAD to check disconnecting the path by both the DTE and the parity in the datastream from the asynchronous PAD are defined. terminal and/or generate parity in the X.28 specifies procedures for character inter­ datastream to the asynchronous terminal. No change and service initialization between an asyn­ parity checking or generation, parity checking, chronous terminal and a PAD. Characters sent and or parity generation are selectable. received must conform to IA5. The PAD transmits • Page wait-allows the PAD to suspend trans­ and expects to receive only eight-bit characters. mission of additional characters to the asyn­ The eighth bit, the last bit preceding the stop ele­ chronous terminal after the PAD has ment, is used for parity checking. transmitted a specified number of line-feed X.28 describes the action the PAD takes characters. when the value of parameter 21 (X.3, parity treat­ ment) is set to 0, 1, 2, or 3. If parameter 21 is set to Recommendation X.28 0, the PAD inspects only the first seven bits and CCITT Recommendation X.28, titled DTE/DCE ignores the eighth bit. When parameter 21 is set to Interface for a Start/Stop Mode Data Terminal 1, the PAD treat~ the eighth bit of the character as Equipment Accessing the Packet Assembly/ a parity bit and checks this bit against the type of Disassembly Facility (PAD) in a Public Data Net­ parity-odd, even, space (0), or mark (I)-used work Situated in the Same Country, describes the between the PAD and the asynchronous terminal. interfacing procedures that allow the PAD to be If it is set to 2, the PAD replaces the eighth bit of connected to an asynchronous terminal. X.28 cov­ the characters to be sent to the terminal with the ers four areas: bit that corresponds to the type of parity used be­ tween the PAD and terminal. When the value is set • Procedures to establish an access information to 3, the PAD checks the parity bit for characters path between an asynchronous terminal and a received from the asynchronous terminal and gen­ PAD; erates the parity bit for characters to be sent to the • Procedures for character interchange and ser­ asynchronous terminal (as in values 1 and 2). vice initialization between an asynchronous ter­ Once the access information path is estab­ minal and a PAD; lished, the asynchronous terminal and the PAD • Procedures for the exchange of control informa­ exchange binary 1 across the interface. This places { tion between an asynchronous terminal and a the interface in the active link state (state 1). When PAD; and the interface is in the active link state, the DTE transmits a sequence of characters that indicates

@ 1991 McGraw-Hili, Inccrporated. Reproduction Prohibited. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA 6 2720 CCITT Packet Switched Data Networking Networking Standards Standards XSerl.. service request (state 2) and initializes the PAD. DTE; they transmit call progress signals. acknowl­ The service request permits the PAD to detect the edge PAD command signals. and transmit operat­ data rate, code, and parity used by the asynchro­ ing information of the PAD to the terminal. Either nous terminal (DTE) and to select the initial pro­ the PAD or the terminal can transmit the break file of the PAD. The service request may be signal. It provides signaling without losing charac­ bypassed, if the terminal is connected to the PAD ter transparency. The prompt PAD service signal via a leased line and the PAD knows the speed, indicates the PAD's readiness to receive a PAD code, and initial profile of the terminal or if a de­ command signal. fault value is used. After the request service signal The temporary storage of characters in an is transmitted, the DTE transmits binary 1, which editing buffer provides editing functions in the places the interface in the DTE waiting state (state PAD. These functions permit the asynchronous 3A). If parameter 6 (X.3, control of PAD service terminal to edit characters input to the PAD before signals) is set to 0, the interface immediately enters the PAD processes them. They include character the PAD waiting state (state 5) after receipt of ser­ delete, line delete, and line display. Character de­ vice request. If parameter 6 is set to other than 0, lete removes the last character in the editing; line the PAD transmits the PAD identification PAD ser­ delete removes the contents of the editing buffer. vice signal (indicates PAD and port identity; is net­ Line display causes the PAD to send a format ef­ work dependent), and the interface enters the fector followed by the contents of the editing buffer service ready state (state 4). The DTE then trans­ to the terminal. mits a selection PAD command signal (state 6), and Procedures for the exchange of user data be­ the PAD transmits an acknowledgment PAD service tween an asynchronous terminal and a PAD apply signal, followed by binary 1, which places the inter­ during.the data transfer state. The values of the face in the connection-In-progress state (state 7). parameters set in X.3 determine which characters If parameter 6 is 0, the PAD will not transmit are transmitted during the data transfer state. For PAD service signals. In this case, the interface is example, if parameters 1 (PAD recall using a char­ placed in the connection-in-progress state after re­ acter), 12 (flow control of the PAD by the start/ ceipt of a valid selection PAD command signal. stop mode DTE), 15 (editing), and 22 (page wait) Once the DTE receives the PAD service sig­ are set to 0, any character sequence may be trans­ nal (state 8) or a sequence of signals in response to mitted by the asynchronous terminal for delivery a PAD command signal, the interface is placed in to the remote DTE during the data transfer state. either the PAD waiting state (state 5) or the data User data is sent to the asynchronous termi­ transfer state (state 9). nal in octets (eight-bit characters) at the appropri­ A fault condition exists if a valid service re­ ate transmission rate for the asynchronous quest signal is not received by the PAD within a terminal; the start/stop bits are added to the data selectable number of seconds after the transmis­ characters. Octets are assembled into packets (see sion of binary 1. If a fault condition occurs, the X.25) and forwarded when enough data has been PAD performs clearing by disconnecting the access received to fill a packet, when the maximum as­ information path. sembly timer delay period has elapsed, when a data The procedures for the exchange of control forwarding character is transmitted, or when a information between an asynchronous terminal break signal is transmitted (parameter 7 is set to and a PAD include PAD command signals, PAD other than 0). service signals, break signals, and prompt PAD ser­ vice signals. PAD command signals flow from the Recommendation X.29 DTE to the PAD; they set up and clear a virtual CCITT Recommendation X.29, titled Procedures call, select a standard profile (PAD parameters) for the Exchange ofControl Information and User that is either CCITT or network defined, request Data Between a Packet Assembly/Disassembly Fa­ current values of PAD parameters, send an inter­ cility (PAD) and a Packet Mode DTE or Another rupt requesting circuit status, and reset a virtual PAD, provides the final step. X.29 describes the call. PAD service signals flow from the PAD to the interfacing procedures that allow the PAD to com­ municate with the X.25 network. It defines the

FEBRUARY 1991 @ 1991 McGraw·HiII, Inccrporated. Reproduction Prohibited. Datepro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Pecket Switched Networking Stenderel. 2720 7 X Se,le. Standards

procedures for the exchange of PAD control infor­ all data packets with the D bit set to O. The DTE mation and the manner in which user data is trans­ may send data packets to the PAD with both the D ferred between a packet mode DTE and a PAD or bit and the Q bit set to 1. When the PAD receives a between two PADS. data packet with both the Q and D bits set to 1. the Recommendation X.29 specifies that control corresponding acknowledgment is transmitted. The information and user data are exchanged between PAD may reset the virtual call if it does not sup­ a PAD and a packet mode DTE or between PADs port the D bit procedure. (Figure 5 shows the bit using the data fields described in X.25. Interface positions for the Q and D bits.) characteristics-mechanical, electrical, functional, The PAD forwards a data packet when a set, and procedural-are also defined as in X.25. read, or set and read PAD message is received or X.29 specifies that the call user data field of when any of the conditions listed in X.28 exist an incoming call or call request packet going tot (e.g., enough data has been received to fill a packet, from the PAD or the packet mode DTE must con­ the maximum assembly timer delay period has sist of protocol identifier and call data fields. A call elapsed, or a data forwarding character is transmit­ request packet need not contain a call user data ted). The PAD never forwards an empty data field to be accepted by the PAD. If the call user packet. data field is present, the PAD transmits it, un­ By sending a set, read, or set and read mes­ changed, to its destination. sage to the PAD, one can read and change the cur­ A call data field's octets consist of user char­ rent values of PAD parameters. Upon receipt of acters sent from the DTE to the PAD during the one of these messages, the PAD delivers to the ca-II establishment phase. This field is limited to 12 DTE any previously received data before it acts on octets. The octet's bits are numbered 8 to 1; bit 1, the PAD message. The PAD responds to a read or the low order bit, is transmitted first. set and read PAD message by sending a parameter Bits 8, 7, 6, and 5 of octet 1 of a user data indication PAD message, which contains a parame­ field of complete packet sequences are the control ter field listing parameter references and current identifier field. This field, which consists of four values. Set allows the changing of parameters. octets, identifies the facility to be controlled. The X.29 also discusses invitation to clear proce­ control identifier field coding for messages to con­ dures, which are used to request that the virtual trol a PAD for an asynchronous terminal is 0000. call be cleared by the PAD; interrupt and discard When the control identifier field is set to 0000, bits procedures, which are used to indicate that the 4, 3, 2, and 1 of octet 1 are defined as the message asynchronous terminal has requested that the PAD code field, which is used to identify specific types discard received user sequences; reset procedures, of PAD messages. as defined in X.25~ error handling procedures by the User sequences perform data exchange. They PAD, which define the actions to be taken by the are transferred in the user data fields of complete PAD when errors are detected; and procedures for packet sequences with the Q bit set to O. Only one inviting the PAD to rese/ect the called DTE (op­ user sequence exists per complete packet sequence. tional), which are used by a packet mode DTE to The PAD transmits all data packets with the D bit request that the PAD clear the virtual call. set to O. The DTE sends a data packet to the PAD with the D bit set to 1. When the PAD receives a data packet with the D bit set to 1, it sends the cor­ X.21 Interface Specifications responding acknowledgment. The PAD may reset The trends in communications engineering lean the virtual call, if it does not support the D bit pro­ toward all-digital networks and the integration of cedure. voice and data. Prospective users of these digital, Control information is exchanged via PAD integrated networks are concerned about an inter­ messages, which contain a control identifier field face that can provide access to voice, data, video and a message code field that may be followed by a teleconferencing, and related services. Currently, a parameter field. PAD messages are transferred in wide variety of connectors, electrical standards, the user data fields of complete packet sequences and user procedures for various services and net­ with the Q bit set to 1. Only one PAD message ex­ works exists-leading to almost insurmountable ists per complete packet sequence. The PAD sends technical and economical problems. Therefore, it is

@ 1991 McGraw-Hili, Incorporated. Reproduction Prohibned. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA CCITT Pack•• Switched Data Networking 8 2720 Networking Standards Standards X Seri••

likely that standards organizations will develop a X.21 has some shortcomings. It does not per­ universal service access interface. Although it would mit the transmission of control information during require certain extensions. X.21 is currently the data transfer. Also. it precludes the insertion of most likely to become a future standard for a uni­ data encryption hardware between the DTE and versal interface in distributed system implementa­ the DCE. Another drawback is the need to modify tions. the DTE/DCE master/slave protocol techniques CCITT Recommendation X.21. Intel/ace Be­ and to supply special crossover cables to facilitate tween Data Terminal Equipment (DTE) and Data DTE-to-DTE or DCE-to-DCE interconnection. Circuit-terminating Equipment (DCE) for S.vnchro­ X.21 uses a different interfacing technique nous Operation on Public Data Networks. defines than that which is normally associated with the physical characteristics and control procedures physical-level interfaces. Instead of assigning each for an interface between DTEs and DCEs. function a specific pin on the connector (e.g .• X.21 is the designated interface for CCITT CCITT V.24 and EIA RS-232-C). X.21 assigns Recommendation X.2S, a packet-switching proto­ coded character strings to each function. col. X.21 can also be used in a non-packet switched The following is a summary of the X.21 stan­ environment. At least two X.21-based public data dard. including the functional descriptions of the circuit switched networks are currently imple­ interchange circuits. phases of operation. electrical mented, one in Scandinavia and one in Japan. characteristics. and mechanical characteristics. The X.21 standard has not gained wide ac­ ceptance in the United States. The reluctance in Functional Descriptions of Interchange Circuits the U.S. to embrace the X.21 standard is due in Four types of X.21 interchange circuits are de­ part to the firm entrenchment of RS-232-C. An­ fined: Ground. Data Transfer. Control. and Tim­ other factor is the cost of implementing X.21. ing. These circuits, outlined in Table 2. are Since X.21 transmits and interprets coded charac­ described below. ter strings, more intelligence must be built into the Ground and Common Return Circuits­ interface, at a higher cost than traditional pin-per­ include two types of common return circuits. DTE function interfaces. Common Return and DCE Common Return. and Certain characteristics of X.21 should ensure one ground circuit, Signal Ground. a more widespread acceptance in the coming years. Signal Ground (Circllit G) establishes the One immense advantage X.21 has over traditional common reference potential for unbalanced interfaces is its capability to assign an almost un­ double-current interchange circuits. If required. it limited number of functions, because there are no reduces environmental signal interference. functional boundaries associated with connector Lowering signaling rates may require two size. Also, X.21 offers a much more sophisticated common return conductors. In this case, two cir­ level of control over the communications process. cuits, DTE Common Return (Circuit Ga) and DCE Another important feature of X.21 is its inherent Common Return (Circuit Gb), are necessary. For a dialing functions, including the provision for re­ further explanation of these circuits, see the Elec­ porting the reasons why a call was not completed. trical Characteristics section of this report. This eliminates the need for a separate data call Data Transfer Circuits-include two Trans­ interface, such as RS-232-C's companion RS-366- mit and Receive data transfer circuits. A, and in switched facilities, results in improved Transmit (Circuit T) transfers signals from response times. the DTE to the DCE during the data transfer Another important aspect is its relationship phase. It also transfers call control signals to the to X.2S. As the packet-switching technique be­ DCE during call establishment and other call con­ comes more widely implemented, the demand will trol phases. be greater for equipment to meet the X.21 stan­ Receive (Circuit R) receives signals transmit­ dard. Internationally, the combination ofX.21, the ted by the DCE from a remote DTE during the ISO HDLC protocol, and X.2S has been used to data transfer phase. This circuit also transfers call form an effective communications path. Another control signals from the DCE during the call estab­ boost for X.21 is IBM's recognition. lishment and other call control phases.

FEBRUARY 1991 @ 1991 McGraw·HiII, Incorporated. Reproduction Prohibited. Datapro Information Services .Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched Networking Standard. 2720 X Serle. Standards •

Control Circuits-include Control and Indi­ Table 2. CCITT X.21 Interchange cation circuits. Circuits Control (Circuit C) transmits signals that con­ trol the DCE for a particular signaling process. The representation of this signal requires additional Direction coding of the Transmit circuit, as specified for the Inter­ Name to DCE from DCE Circuit procedural characteristics of the interface. During change Type the data phase, Circuit C remains in the ON condi­ Circuit tion. G Signal Indication (Circuit I) indicates the call control ground or common process to the DTE. The representation of this sig­ return nal requires additional coding of the Receive cir­ Ga DTE x Ground cuit. When Circuit I is on, it signifies that signals common return on the Receive circuit contain information from Gb DeE x the remote DTE. When Circuit I is off, it signifies a common control signaling condition, defined by the Circuit return R bit patterns, as specified by the procedural char­ T Transmit X Data acteristics of the interface. Transfer Timing Circuits-includes Signal Element R Receive X Timing and Byte Timing. c Control X Signal Element Timing (Circuit S) provides Indication X Control the DTE with signal element timing information. s Signal X element For this function, Circuit S turns on and off for timing nominally equal periods of time. B Byte timing X Timing X.21 defines different roles for the DTE and DCE in regard to signal element timing. During the off-to-on transition, the DTE presents a binary sig­ Quiescent Phase-the quiescent phase is the nal on Circuit T and a condition on Circuit C. The period during which the DTE and the DCE signal DCE presents a binary signal on Circuit R and a their capability to enter the call control phase or condition on Circuit I during the off-to-on transi­ the data transfer phase. It is characterized by the tion. The DCE transfers the signal element timing appearance of basic quiescent signals from the across the interface as long as the timing source is DTE and DCE. Various combinations of these qui­ capable of generating this information. escent signals result in different interface states, or Byte Timing (Circuit B) provides the DTE quiescent states. with eight-bit timing information for synchronous There are three DTE quiescent signals. DTE transmission. Use of this circuit is not mandatory. Ready indicates the readiness of the DTE to enter Circuit B turns off whenever Circuit S is in the ON the other operational phases. DTE Uncontrolled condition, indicating the last bit of the eight-bit Not Ready indicates the DTE is unable to enter byte. At all other times within the period of the certain operational phases, usually due to an ab­ eight-bit byte, Circuit B remains on. normal condition. DTE Controlled Not Read.v indi­ cates that although the DTE is operational, it is Phases of Operation temporarily unable to accept incoming calls for The X.21 standard defines four phases of opera­ circuit switched service. tion: the Quiescent Phase, the Call Control Phase, There are two DCE quiescent signals: DCE the Data Transfer Phase, and the Clearing Phase. Ready and DCE Not Ready. DCE Ready indicates Each step of the operational phases places the the DCE is ready to enter operational phases. DCE DTE and DCE in a certain state. See Table 3 for a Not Ready indicates that no service is available; it listing of these states and their associated signals is also signaled whenever possible during network ( on the interchange circuits. fault conditions and during the period when test loops are activated.

@ 1991 McGraw-Hili, Incorporated. Reproduction Prohibited. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA CCI" Packet Switched Data Networking 10 2720 Networking Standard. Standards X Serle.

Table 3. X.21 States: Names, Signalling, and Transitions

Signals on T,C and R,I Circuits Transitions· / Phase of OTE OCE State Opera- to state to state Number Stete Name tion T C R I number number 1 Ready Q 1 OFF 1 OFF 2,13S,14,24 8,13R,18 2 Call request CC 0 ON 1 OFF 3,15 3 Proceed-to-select CC 0 ON + OFF 4,15 4 Selection signal CC IA5 ON + OFF 5

5 DTE Waiting CC 1 ON + OFF 6A,11,12 6A DCE Waiting CC 1 ON SYN OFF 7,10,11,12 6B DCE Waiting CC 1 ON SYN OFF 10 bls, 11,12 7 Call progress signal CC 1 ON IA5 OFF 6A,10,11,12 8 Incoming call CC 1 OFF BEL OFF 15,9 9 Call accepted CC 1 ON BEL OFF 6B,11,12 10 DCE provided information CC 1 ON IA5 OFF 6A,11,12 10 bis DCE provided information CC 1 ON IA5 OFF 6B,11,12 11 Connection in progress CC 1 ON 1 OFF 12 12 Ready for data CC 1 ON 1 ON 13 13 Data transfer DT 0 ON 0 ON 13R 13S,DCE not ready

13R Receive data OT 1 OFF 0 ON 13 1 13S Send data OT 0 ON 1 OFF 7 13 14 DTE Controlled not ready, DCE ready Q 01 OFF 1 OFF 1,24 23 15 Call colliSion CC 0 ON BEL OFF 3 16 DTE Clear request C 0 OFF X X 17 (see Note) 17 DCE Clear confirmation C 0 OFF 0 OFF 21 18 DTE Ready, DCE Not ready Q 1 OFF 0 OFF 22 1 DCE Not ready Q 0 ON 0 OFF 1,13,13S 19 DCE Clear indication C X X 0 OFF 20 (see Note)

20 DTE Clear confirmation C 0 OFF 0 OFF 21 21 DCE Ready C 0 OFF 1 OFF 22 DTE Uncontrolled not ready, DCE not ready Q 0 OFF 0 OFF 18 24

23 DTE Controlled not ready, DCE not ready Q 01 OFF 0 OFF 18,22 14 24 DTE Uncontrolled not ready, DCE ready Q 0 OFF OFF 22

Any state (see Note) X X X X 16 19 • All other transitions are considered invalid. Note: DCE Clear indication or DTE Clear request may be entered from any state except Ready.

Key to Table: o and 1-Steady binary conditions 01-Alternate binary 0 and binary 1 Q-Qulescent Phase X-Any value CC-Cail Control Phase OFF-Continuous off (binary 1) DT-Data Transfer Phase ON-Continuous on (binary 0) "'- .. / T-Transmlt Interchange circuit IA5-Characters from CCITT Alphabet #5 C-Control interchange circuit +-IA5 character 2/11 R-Rece/ve Interchange circuit BEL-IA5 character 0/7 I-Indication Interchange circuit SYN-IA5 character 1/6

FEBRUARY 1991 @ 1991 McGraw-Hili, Incorporated. ReproduciloriProhlbltild. Datapro Information ServiCes Gtoup. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 Networking Stend.reI. 11 X Serle. Standards

There are six quiescent states: • Incoming Call; indicated by the DCE. In re­ sponse, the DTE signals Clear Request, DTE • Ready Uncontrolled Not Ready, or DTE Controlled • DTE Controlled Not Ready, DCE Ready Not Ready. • DTE Ready, DCE Not Ready • DCE Waiting. • DTE Uncontrolled Not Ready, DCE Not Ready • Call Progress Signal Sequence is transmitted by • DTE Controlled Not Ready, DCE Not Ready the DCE to the calling DTE to indicate that cir­ • DTE Uncontrolled Not Ready, DCE Ready cumstances have arisen to prevent the connec­ tion from being established, to report the See Figure I for a diagram of the quiescent states progress made toward establishing the call, or to and the transitions that are allowed between these signal that problems have been detected and states. that the call needs to be cleared and reset. Call Control Phase-the call control phase for • DCE-Provided Information Sequence, transmit­ circuit switched service contains many elements ted from the DCE to the calling DTE. It consists and procedures. Characters used for call control of DCE-provided information blocks, such as are selected from lAS, a seven-bit plus parity inter­ line identification and charging information. national code outlined in CCITT Recommenda­ Line Identification is transmitted by the DCE tion V.3. Each call control sequence to and from to the calling DTE during the DCE-Provided the DCE is preceded by two or more continuous Information state immediately after all call SYN characters. For error checking of call control progress sigl)als, if any, are transmitted. Both characters, odd parity is specified. calling and called line identification are op­ The following elements of the call control tional. Line identification consists of the inter­ procedure are outlined in X.21: events of call con­ national data number, as assigned in CCITT trol procedures, unsuccessful call, call collision, Recommendation X.121, International Num­ direct call, and facility registration/cancellation bering Plan for Public Data Networks. The DCE procedure. These elements are summarized below. transmits Charging Information during the The events of the call control procedures in­ DCE-Provided Information state. It informs the clude the following: subscriber of either the monetary charges for a • Call Request, signaled by the DTE to indicate a call, the duration of the call, or the number of request for a call. units used during the call. • Proceed to Select, used when the network is pre­ • Connection-In-Progress, indicated by the DCE. pared to receive selection information. It is • Ready for Data, transmitted by the DCE when transmitted by the DCE to the DTE within the connection is available for data transfer be­ three seconds of the call request signal. tween DTEs. • Selection Signal Sequence, transmitted by the DTE. A selection sequence consists of a facility An unsuccessful call occurs when a required con­ request block, an address block, a facility re­ nection cannot be established. In this case, the quest block followed by an address block, or a DCE indicates the failure and its reason to the call­ facility registration/cancellation block. A facil­ ing DTE through a call progress signal. ity request block comprises one or more facility A call collision can occur in one of two ways: request signals, which consist of a facility re­ a DTE detects a call collision when it receives In­ quest code containing one or more facility pa­ coming Call in response to Call Request. A DCE rameters. An address block contains one or detects a call collision when it receives Call Re­ more address signals. Address signals consist of quest in response to Incoming Call. When the DCE either a full address signal or an abbreviated detects a call collision, it will indicate Proceed to address signal. select and cancel the incoming call. ( • DTE Waiting.

~ 1991 McGraw-Hili. Incorporated. Reproduction Prohibited. FEBRUARY 1991 Datepro Information Services Group. Oslran NJ 08075 USA 12 2720 COITT Packet Switched Data Networking Networking Standard. Standards X ..rI ••

The DTE indicates a request for a direct call For operation over switched facilities, the by signaling DTE Waiting after receiving the Pro­ DTE may send bits to a corresponding DTE after ceed to Select signal. If necessary, the DTE may receiving the Ready for Data signal. During data choose an addressed call by presenting the correct transfer, control and interchange circuits are in the Selection signal. ON condition, and data is transmitted over the The facility registration/cancellation proce­ transmit and receive circuits. Data transfer may be dure is optional. A facility registration/cancellation terminated by clearing, which is defined below. signal consists of up to four elements in order: fa­ Two basic signals are used for operation over cility request code, indicator, registration parame­ leased, point-to-point facilities. Send Data trans­ ter, and address signal. Not all of these elements mits data by the DTE on Circuit T; the remote are required in the facility registration/cancellation DTE's Receive Data signal receives data over Cir­ signal. Also, a number of these signals may be cuit R. To terminate the data transfer, the DTE linked to form a block. In response to acceptance signal places its transmit circuit in the binary 1 or rejection of the facility registration/cancellation condition. The DCE indicates termination of data action, the network provides the appropriate Call transfer by placing its receive circuit in the binary Progress Signal. 1 condition, its control circuit in the OFF condi­ Data Transfer Phase-when the DTE is in the tion, and its indications circuit in the OFF condi­ data transfer phase, any bit sequence may be trans­ tion. mitted. X.21 defines the data transfer phase for Both the central and remote DTEs use the three types of connections: switched; leased, point Send Data and Receive Data signals for operation to point; and leased, centralized multipoint. over leased, multipoint facilities. The central DTE

Figure 1. Quiescent States

Definjtions; +

n State number t Signal on T circuft c Signal on C circuit r Signal on R circuit Signal on I clrcu~

Transition with indication DCE DCE DCE • of whether DTE or DCE is responsible for trans~ion

, / The above diagram indicates transitions that are allowed in X.21 networks. Other transitions are possible and may be allowed in some networks. See Table 3 for a listing ofpossible transitions.

FEBRUARY 1991 @) 1991 McGraw-Hili, Incorporated. Reproduction Prohibited. Datapro Information Services Group. Delran NJ 08075 USA Data Networking CCI" Packet Switched Networking Slllnd.rd. 2720 13 XS.rI.. Standards

delivers data transmitted to all remote DTEs; re­ ISO 4903 assigns connector pin numbers to a mote DTEs (one at a time) transmit data to the 15-pin interface between DTE and DCE equip­ central DTE. A remote DTE may send data to the ment. Table 4 presents a chart of these X.21 pin central DTE while the central DTE is sending to all assignments as they relate to the X.26 and X.27 remote DTEs. standards. Clearing Phase-either the DTE or the DCE may initiate clearing. The DTE indicates its desire X.2181s to enter the clearing phase by transmitting DTE Although X.21 is the specified interface for X.2S, Clear Request. The DCE responds by signaling alternative interfaces also exist. One of these is DCE Clear Confirmation, followed by DCE Ready. X.21 bis. Clearing by the DCE takes place when it CCITT Recommendation X.21 bis, the physi­ transmits DCE Clear Indication. The DTE re­ cal and functional equivalent to CCITT V.24, de­ sponds with the DTE Clear Confirmation signal, fines 25 interchange circuits between DTEs and followed by the DCE signaling DCE Ready. DCEs. CCITT V.24 is compatible with EIA Stan­ dard RS-232-C. The X.21 bis recommendation, Electrical Characteristics accepted as the interim interface for X.25, will be .?'.21 uses two types of electrical characteristics, gradually replaced by X.21 as more equipment is each for different system requirements. manufactured to meet X.21 specifications. For synchronous operation at 9600 bps and below, the interchange circuits at the DCE side of the interface must comply with CCITT Recom­ Recommendation X.25 mendation X.27. The DTE side can comply with The development of Recommendation X.25 was either X.27 or another CCITT Recommendation, stimulated by the need for a standard interface be­ X.26. For synchronous operation at signaling rates tween the packet-switching networks already devel­ above 9600 bps, interchange circuits at both the oped or being developed by many industrial DTE and DCE sides of the interchange circuits nations and by the requirement that no terminal must comply with X.27. equipment be denied access to packet switched ser­ X.26 is defined in CCITT Standard V.IO. It vices. describes the electrical characteristics for unbal­ X.25 is a dy"namic standard with many varia­ anced interchange circuits. X.26 calls for both a tions in the U.S. and abroad. Currently, X.25- DTE and DCE common grounding arrangement. based packet switched networks exist in Australia, The maximum suggested cable length is 1,000 Austria, Belgium, Canada, France, Ireland, Ger­ meters, and the maximum data rate is lOOK bps. many, Hong Kong, Italy, Japan, Mexico, the Neth­ X.27 is defined by CCITT Standard V.II, erlands, Portugal, Singapore, the Soviet Union, which describes the electrical characteristics for South Africa, Spain, Switzerland, the United King­ balanced operation. Maximum suggested cable dom, and the United States. Since X.2S is a dy­ length is 1,000 meters, and the maximum data rate namic standard with many extensions and optional is 10M bps. features, these networks are not totally compatible with one another. Those located in Europe have ..echanlcal Characteristics the highest level of mutual compatibility. The mechanical characteristics for X.21 are out­ Since the establishment of X.25, additional lined in the ISO Standard 4903, approved by the user-level protocols have been developed. These International Organization for Standardization protocols provide the interfaces between different (ISO). The standard, entitled Data types of terminals and the X.25 interface. X.3, Communication-I5-pin DTEIDCE Interface Con­ X.28, and X.29, informally called the Interactive nector and Pin Assignments, was published in June Terminal Interface (ITI), were the first of the pro­ 1980. tocols to interface to X.25. They relate to the sup­ port of asynchronous, low-speed terminals by ( packet switched networks. These are logical com­ plements to X.25 because they permit specific sets

CI 1991 McGraw-Hili, Incorporated. Reproductlon Prohlblted. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA 2720 CCITT Packet Switched Data Networking 14 Networking Standards Standards X S.rl••

of terminals to interface to the packet networks supports X.32 with regard to the public switched using the X.25 interface. telephone network or a circuit switched public data The X.25 interface standard provides for the network. dial-in and dial-out access. backup for connection of terminals and computers to public leased line connections. long-distance access to the packet-switching networks. X.25 outlines three lay­ network. and teletex. ers of operation: the Physical Layer. the Link Datagram was deleted in 1984. while the fol­ Layer, and the Packet Layer. These layers parallel lowing packet-level services were made available as the bottom three layers of the ISO Reference options: Model for Open Systems Interconnection. The • Registered Private Operating Agent (RPOA) Se­ Physical Layer calls for CCITT X.21 as the physi­ lection permits the use of one or more networks cal and electrical interface but accepts X.21 bis. a to route a call to its destination. If the user se­ functional equivalent of RS-232-C, as an interim lects only one network. either the basic or ex­ standard. The Link Layer uses the procedures of tended format of the RPOA Selection can be the HDLC protocol standard. The Packet Layer used; if more than one network is chosen. the defines procedures for constructing and controlling extended format is used. a data packet. The 1984 revision of Recommendation X.25 • Call Redirection permits the rerouting of calls if added specifications for X.21 access and expanded the first tried route fails. the potential of packet operations, allowing users • Call Redirection Notification informs the recipi­ to actively gain access to the X.25 port, identify ent of the forwarded call that the call has been themselves, and validate their connection through redirected. passwords. This change reoriented the X.25 stan­ • Called Line Address Modified Notification tells dard toward switched access through both X.21 the caller, within a call confirmation packet. facilities and the public telephone network. It now that the call has been redirected. • Hunt Group distributes incoming calls that have Table 4. X.21 Pin Assignments an address associated with the hunt group. • Charging In/ormation gives the caller informa­ tion on time and charges and requires a new Pin Employing Employing field in the call-clearing packet format. Number X.26 X.27 See Note (3) See Note (3) • Local Charging Prevention is a security facility 9 Ga T(B) that prevents reverse or third-party call charges. 2 T T(A) • Network User Identification accommodates user 10 Ga C(B) ID, billing, and on-line facilities registration. 3 C C(A) This permits users to communicate directly 11 R(B) R(B) with the packet data network to change the pa­ 4 R(A) R(A) rameters of their subscriptions. 12 I(B) 1(8) 5 I(A) I(A) The packet level is an octet-oriented (eight bits per 13 S(B) S(B) octet) structure. Packet sizes can vary from 1,024 6 SeA) S(A) to 2,048 octets, but only within a network. 14 B(B) B(B) Network-to-network exchange is limited to 128 7 B(A) B(A) octets. Closed user group facilities can accommo­ 15 Reserved for Reserved for date very large private-packet networks, although future use future use the number of closed user groups to which a DTE 8 G G can belong is network dependent. Notes: Link-level changes implemented in 1984 cre­ (1) X.21 pin assignments are defined by ISO 4903-1980. (2) Where balanced cirults are, the associated pairs are desig- ated a clear separation between the Link Access nated "A" and "B." Procedure (LAP) and the Link Access Procedure (3) Pin 1 Is reserved for connection of the shield or shielded In- terconnecting cable. Balanced (LAPB). Multilink procedures for a single interface were also implemented. LAPB underwent

FI:13RUAAY 1991 @ 1991 McGraw-Hili, Incorporated. ReproduGtlon PrOhibited. Datapro Information Services Group. Delran NJ 08015 USA Data Networking CCITT Packet Switched 2720 15 Networking Standard. X Serle. Standards

an alignment with the single-link procedure of network-operated PAD. discussed earlier in this X.75. An extended numbering option (modulo report. Terminal transmissions are stored in a 128) was added to LAPB to enable the sequencing buffer at the PAD. There they are assembled into of 127 frames and the use of satellite facilities. In packets and sent to the device at the other end of addition, the 1984 revisions to X.25 refined the the virtual circuit. Packets arriving at the PAD procedure for the implementation of the D-bit. from the network are reassembled into the appro­ polished technical accuracy, and defined the rules priate format before they are sent on to the termi­ for new fields and formats. nal. The PAD is programmed and configured to interface properly with the protocol and physical X.2S Communications characteristics of the user's device. Data presented X.25 is titled Interface Between Data Terminal to the PAD is reformatted into X.25 format and Equipment (DTE) and Data Circuit-terminating forwarded to the DCE. Equipment (DeE) for Terminals Operating in the Recommendation X.25 is divided into those Packet Mode on Public Data Networks. It provides specifications required for a device or network to a precise set of procedures for communications comply and those that are optional. Approximately between DTE and DCE for terminal equipment one third are required: the remaining two thirds of operating in a packet environment. The DCE in the specifications are optional. this case is a node processor that serves as an The excess throughput capacity inherent in entry/exit point on the packet network side of the the X.25 standard allows for future network growth user/network interface. and technological progress. For example. a single The Data Terminal Equipment (DTE) is a X.25 interface can theoretically handle 4.095 vir­ programmable device on the user side of the user/ tual channels, packet sizes up to 2.048 bytes each. network interface. The DTE is located at the user and packet sequencing up to 128 packets per logi­ site when the on-site equipment supports X.25; at cal channel. Most network suppliers' nodal proces­ such installations, the DTE can be either a com­ sors are too small to handle this much traffic puter, a front-end processor, or an intelligent ter­ through a single interface. Therefore. in practice. minal, as shown in Figure 2. The DTE can be a the support offered over each interface is limited to group of intelligent terminals (multiplexed to avoid the current capacity of the network's access node. the use of multiple lines) that transmit data over the packet network to a remotely located host. It Functional Layers can also be a processor acting on calls received Recommendation X.25 defines three functional from multiple locations that communicate over the layers: the Physical Layer (Level I). the Link Layer packet network. (Level 2), and the Packet Layer (Level 3). These are Regardless of the device or application, all consistent with the first three layers of the ISO Ref­ DTEs present standard formatted data and control erence Model for Open Systems Interconnection information to the DCE over standard communi­ (OSI). (Although OSI labels its third layer the Net­ cations facilities. Devices operate over the network work Layer, its function parallels that of the Packet in the virtual circuit mode. Essentially, the user Layer. Both provide the means to establish. main­ causes the network to establish a logical circuit tain, and terminate connections.) connection with the receiving station for the trans­ Levell: Physical Level-outlines the physical. mission of multiple contiguous packets. (The ac­ functional. and electrical characteristics of the tual physical circuit over which individual packets DTElDCE interface. X.21 uses the transmission of are transmitted can vary during a session, but the coded character strings across a 15-pin connector logical circuit ensures presentation of each packet to define standard interface functions. e.g., Trans­ to the receiving station in the proper order.) Infor­ mit Data. Its specifications are defined in CCITT mation delivery is so rapid that the user appears to Recommendation X.21. have an end-to-end, dedicated channel. Although X.21 is the recommended physical­ Users who do not support X.25 can access the level interface for X.25, the availability of data public data network for packet data transmission; communications equipment with X.21 capabilities however, they cannot transmit directly to a DCE or network node. They must transmit to a special

@ 1991 McGraw-Hili. Incorporated. Reproduction Prohibited. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA f8 2720 CCITT Packet Switched Data Networking Networking Stllndard. Standards X S.rIe.

Figure 1. A Sample User Terminal Configuration for Operating on Q Public Data Network (PDN) in Packet Mode, with X.25 as the Interface to the Network

User Location PubHc PacJcet-Switching Network Virtual Circuit OTE DCE Computer with X.25 Level 2 I Modem I and L8IIel3 - /" Modem - Software - I I Node Processor I I with X.25 OTE I I Level 2 I and I I Level 3 Front- Software Host End Computer Modem I /" I Modem Processor - - - I - I X.25 Level 2 and 3 Software I

OTE I Intelligent Terminal with I X.25 Level 2 - I and Level 3 - I Software I ~ I I i I I ~ /" I Modem I I :t= - OTE I =::I I :::E I InteUigenl Terminal whh I I X.25 Level 2 ~ and Level 3 - I ~ Software r--

• Signifies an interlace thai must conform to X.25 Level 2. Physical Interlace Standards. is limited, especially in the United States. As a re­ and DCE can conduct two-way, simultaneous. in­ sult, the CCITT has accepted an interim recom­ dependent transmissions. The procedures use the mendation, X.21 bis, as the X.25 physical principles and terminology of the High-level Data interface. X.21 bis is functionally equivalent to Link Control (HOLC) protocol specified by the EIA RS-232-C, which assigns a separate function International Organization for Standardization. to each pin across a 25-pin connector. Level 2 comprises three procedures: the Link Level 2: Link Level-describes the link access Access Procedure (LAP), the Link Access Proce­ procedures used for data interchange between a dure Balanced (LAPB), and the Multilink Proce­ DCE and a DTE. In the CCITT Recommendation dure (MLP). The original X.25 data link control X.21, International User Classes ojService in Pub­ procedure embraced LAP, which is similar to the lic Data Networks, X.25 specifies that the DTE and HOLC Asynchronous Response Mode (ARM). DCE must operate in the following user classes of Due to some incompatibilities with some elements service: class 8 (2400 bps), class 9 (4800 bps), class of procedures, Level 2 of the X.25 Recommenda­ 10 (9600 bps), or class 11 (48K bps). The proce­ tion was revised to incorporate LAPB. Similar to dure calls for a full-duplex facility so that the DTE the Asynchronous Balanced Mode (ABM) of

FEBRUARY 1991 @ 1991 McGraw-Hili, InCorporated. Reproduction Prohibited. Datapro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 17 Networking Standards xSeri.. Standards

HDLC, it provides for a "balanced" configuration; The Supervisory Frame transmits one of that is, each side of the link consists of a combined three supervisory commands and cannot contain primary/secondary station. The Multilink Proce­ user data. The Control field of the S-Frame con­ dure is used for data transmission on one or more tains the command, the P/F bit, and an N(R). single links. Each link conforms to the X.25- Table 6 summarizes the commands and responses defined frame structure and to the elements of pro­ possible in the Control field. cedure described in LAPB. The Unnumbered Frame extends the number Software in both DTE and DCE performs oflink control functions, without incrementing the Level 2 processing. This software appends control sequence counts at either the sending or receiving information onto packets that are ready for trans­ station. mission, maintains control of the transmissions, A U-Frame control field contains the Un­ performs transmission error checking, and strips a numbered command and the P/F bit. One of the successfully received frame down to a packet. The U-Frame commands initializes the DTE/DCE net­ packet consists of packet control information and work to the Asynchronous Response Mode (ARM) (optionally) user data; packets are discussed in de­ of operation, which permits both DTE and DCE to tail later in this report. initiate transmission to each other. HDLC specifies certain control fields that The Frame Check Sequence (FCS) field con­ must be appended to both ends of a packet, result­ tains a 16-bit check pattern created at framing time ing in a transmission format called a/rame. Ap­ by generating check bits based upon the binary pended in front of a packet are a beginning Flag value of the data in the packet. The receiving de­ field, an Address field, and a Control field. Ap­ vice performs the same calculation and compares pended behind the packet are a Frame Check Se­ its results with the value that the sending device quence field and a closing Flag field. Figure 3 gives had placed in the FCS field. If these values do not the size and description of each field. The receiving agree, the frame has a transmission error and is device uses the two Flag fields to ascertain the be­ discarded. Discarding causes the receiving device ginning and ending of a frame. A single Flag field to send an S-Frame with the Reject Recovery com­ can be used as the closing flag for one frame and mand. This S-Frame contains the frame numbered the opening flag of the next frame. The Address N(R), thus acknowledging receipt of all frames field, under HDLC, is used to identify the sta­ numbered N(R)-1 and below, and initiates retrans­ tion(s) for which the command is intended in com­ mission of frames N(R) and above. mand frames or to identify the station sending the In effect, Level 2 envelops the packet in con­ response in response frames. trol information and prescribes procedures that The Control field identifies the type of frame ensure a high degree of transmission accuracy and and supplies control information pertinent to that detection of lost frames during transmission. type of frame. A frame can be either an Informa­ Level 3: Packet Level-describes the packet tion Frame (I-Frame), a Supervisory Frame (S­ format and control procedures for the exchange of Frame), or an Unnumbered Frame (U-Frame). See packets between the DTE and the DCE. In addi­ Table 5 for the encoding of this field. tion to the user data packet format, there are sev­ An Information Frame contains a user eral formats for DTE/DCE administrative packet. The Control field of the I-Frame contains messages. the N(S) transmitter Send Sequence Count, the The software for formatting packets and con­ N(R) transmitter Receive Sequence Count, and the trolling packet exchange is the Level 3 software Poll/Final (P/F) bit. N(S) is the sequence number resident in both the DTE and the DCE. Figure 4 of this frame sent from this transmitter to this re­ shows the relationship of Level 3 and Level 2 soft­ ceiver. N(R) is the sequence number ofthe next ware operating in the DTE and the DCE. In the frame this transmitter expects to get from the re­ DTE, a user application program normally pre­ ceiver when the receiver becomes a transmitter. sents the data to be transmitted to the operating The Poll/Final bit indicates whether a transmission system's communications access method. The ac­ ( acknowledgment is required or when the final cess method invokes Level 3 software to append frame in a stream has been transmitted. the packet header information, then invokes Level

@ 1991 McGraw·HIII. Incorporated. Reproduction Prohibited. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA COITT PlICket Switched Data Networking 18 2720 Networldntl Standard. Standards X ......

Figure 3. Frame Formats for Packet Mode Transmission /

~I"~------Framel------~"~I

I...~.... -----packet----""~~I

Packet Field Control Flag Address Control User Data Sequence Flag Field Infomation Check

Fields Generated by:

Application Program

Level 3 Software

Level 2 Software Size, Field !!!!! Description 8 Value is 01111110. Flag 8 Value for DTE to DCE command frame is "10000000". Address Value for DTE to DCE response frame is "11000000":

Value for DCE to DTE command frame is "11000000". Value for DCE to DTE response frame is "10000000".

8 I Frame: Control Bit 1 is "0". Bits 2, 3, 4, are N(S) (transmitter send sequence count). Bit 5 is PIF (PoIVFinal) bil Bits 6, 7, 8 are N(R) (transmitter receive sequence count).

S Frame: Bit 1 is "1". Bi12 is "0". Bits 3, 4 identify Supervisory Command. Bit 5 is P/F (PoIVFinal) bit. Bits 6,7,8 are N(R) (transmitter receiver sequence count).

U Frame: Bit1is1. Bit2is 1. Bits 3, 4 are first part of Unnumbered Frame Command identifier. Bit 5 is PIF (PoIVFinal) bit. Bits 6, 7,8 are second part of Unnumbered Frame Command identifier.

Packet 24 Control information. Control 1,024 1,024 data bits are normal maximum; exceptional maximum is 2,040 data bits User Data 16 Check bits for user data. Field Sequence Check 8 Value is "01111110". Flag

2 software to create the frame. Packet header for­ detection and sequence checking and to strip the mats are discussed below. accepted frame down to a packet. The packet is Once the frame is created, it is ready for presented to Level 3 software for packet-level pro­ transmission. Upon receiving the frame, the receiv­ cessing and to prepare the packet for transmission ing DCE invokes Level 2 software to perform error to the destination DCE. The physical address of

FEBRUARY 1991 C> 1991 McGraw-Hili, I~ated. Reproduction Prohibited. Datapro Information Servioas Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 Networkln. Standards 11 X ..rle. Stendards

Table 5. X.25 Level 2 Commands and Responses

Commands (1) Rasponses Encoding of Control Field Format 1 2 3 4 5 8 7 8 Information (information 0 N(S) P N(R) transfer Supervisory RR (3) (receive ready) RNR (receive not ready) 1 0 0 0 P/F N(R) RNR (3) (receive not ready) RNR (receive not ready) 1 0 1 0 P/F N(R) REJ (3) (reject) REJ (reject) 1 0 0 1 P/F N(R) Unnumbered SARM (set asynchronous OM (disconnected (2,3) response mode) (2) mode) P/F 0 0 0 SABM (set asynchronous (2) balanced mode) 1 1 P 1 0 0 DISC (disconnect) 0 0 P 0 0 UA (unnumbered acknowledgment) 0 0 F 1 1 0 CMDR (command reject) 0 F 0 0 FRMR (frame reject) Note 1: The need for, and use of, additional commands and responses are for further study. Note 2: OTEs do not have to implement both SARM and SABM, furthermore OM and SABM need to be used if SARM only is used. Note 3: RR, RNR and REJ supervisory command frames are not used by the OeE when SARM is used (LAP).

the destination DTE, derived from the call initia­ pass the user information to the appropriate appli­ tion, is inserted into the packet during Level 3 pro­ cation program via the communications access cessing. At the network destination DeE, Level 3 method. Table 7 summarizes the types of packets software reformats the packet control information exchanged between terminals. and presents the resulting packet to Level 2 soft­ The Packet Header consists of three octets. In ware; Level 2 software includes the packet in a Octet I, the first four bits represent the Logical frame for transmission to the destination DTE Channel Group Number, and the final four bits (user). At the destination DTE (user), Level 2 soft­ represent the General Format Identifier. Octet 2 ware performs error detection and sequence check­ represents the Logical Channel Number, and Octet ing and strips accepted frames down to the packet. 3 represents the Packet Type Identifier. See Level 3 software is then applied to the packet to Figure 5. strip the header information from the packet and

Table 6. Summary of Packets Exchanged between Terminals during a Virtual Call

Events Activity at Calling DTE Activity at celled DTE Call Initiation Call Request packet is sent Incoming Call packet Is received Call Accepted packet is sent Call Connected packet is received

Data Transfer Data packet sent Data packet received Ready Receive packet sent Ready Receive packet received Data packet B sent" Data packet A sent" Data packet A received" Data packet B received"

Disconnect Clear Request packet sent Clear Indication packet received Clear Confirmation packet sent ( Clear Confirmation packet received "Two-way (full-duplex) transmission of data packets between terminal equipment.

@ 1991 McGraw-Hili, Incorporated. Reproduction Prohibited. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA 20 2720 cern Packet Switched Data Networking Networking Stand...... Standards X Serte.

The Logical Channel Group Number and the when the user subscribed to the network. The Call Logical Channel Number provide the routing infor­ Request packet also informs the network of the mation that directs the packet over one of the logi­ calling DTE's address and of the called DTE's ad­ cal channels. The numbering system for the logical dress. Until the call is disconnected. the network channels is dynamic and refers to a switched data retains the addresses of both devices associated path within the packet network. with the logical channel number. Therefore. each The General Format Ident(fier indicates the data packet that is transmitted needs to contain general format of the rest of the header. Specific bit only the logical channel number. The network ap­ patterns are established for the following: call setup pends the destination address just before routing packets; clearing. flow control. interrupt. reset. re­ the packet. At the time the Call Request is issued. start. and diagnostic packets; and data packets. the logical channel indicated by the calling DTE The Packet Type Ident(fier establishes the must be in a Ready state. that is. not being used to packet's function. Examples of functions include handle another call. Upon receipt of the Call Re­ call setup. priorities. and data. If applicable. it may quest by the called DTE, the specified logical chan­ identify the packet's place in sequence. See Table 7 nel is designated as being in the DTE Waiting for the various packet types. state. The packet is then transmitted to the destina­ tion DCE. Packet-Level Procedures The destination DCE transmits an Incoming X.25 Level 3 defines procedures for call initiation, Call packet to the originating DTE and places the data transfer, interrupts, reset, restart, and clear­ logical channel in the DCE Waiting state. ing. Some of these procedures are summarized be­ The called DTE indicates its willingness to low. accept the call by transmitting a Call Accepted Call Initiation: The Level 3 software in a packet across the DTE/DCE interface. The Call DTE initiates a transmission by sending a Call Re­ Accepted packet specifies the same logical channel quest packet. This packet enables the calling DTE that is indicated by the Incoming Call packet. This to request the opening of a logical channel. The places the specified logical channel in the Data calling DTE designates the channel number based Transfer state. (The logical channel at the calling upon the original assignments that were made DCE is still in the Wait state.)

Figure 4. X.25 User and Network Software Relationships

t------Uoer Location----~f------:Communication.----+_----Pubic Packet· Faciily Switching Network

DeE OTE Network Nod. Procesaor

r----'-----I I I I I I I Communication. Ac ..s. Method I I U.. r r----I r----I I I I I To Program I US I I US I L_I3, Le ..l2, I X.2S Level 2, X.2S Leyel 3. I Other I Packet Level I I Frame Level I I Frame L... I Packet Level I I Network ""- ____1 I Nodes '- -- --' I I I I I I I L,~~, I Phrsicsllnt.~ace I

LX.25Le.. 11. Phroicallnte~a..

FEBRUARY 1991 @ 1991 McGraw-Hili, Incorporated. Reproduction Prohibited. Oatapro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 21 Networking Standard. X Serle. Standards

Figure 5. X.25 Packet Header Format

1..... ,...--IOctet 1 -----7I*~--Octet 2--~*~--Octel 3--1

D-BII Q·Bit M-B.

Bits 1 2 3 4:5 6 7 8 1 2 3 4 5 678 1 2 3 4 5 6 7 8 Logical General Channel Fomat Logical Channel Number Packet Type Identifier Group Idenifier Number

A Call Connected packet is then sent by the packet to the sending device or deny authorization DCE to the calling DTE and sets the logical chan­ to transmit by issuing a Receive Not Ready packet. nel status to the Data Transfer state; the logical When transmitting Data packets, the Inter­ channel is then ready for transfer of data packets. rupt packet can bypass flow control (authorization This applies only for virtual circuit connections; procedures), A DTE can transmit an Interrupt for permanent virtual circuit connections, the as­ packet to another DTE. To avoid swamping the signed logical channels are always in the Data receiver with Interrupt packets, the sender cannot Transfer state. send a second Interrupt packet until the first Inter­ It is possible for a DCE to receive a Call Re­ rupt packet is acknowledged by an Interrupt Con­ quest (from a DTE) and an Incoming Call (from firmation packet. the network) simultaneously, with both messages While Data packets or Interrupt packets are specifying the same logical channel. This is called a being transmitted, issuing a Reset Request packet call collision; when a collision occurs, the DCE can reset the call. A DCE initiates a reset by issuing cancels the Incoming Call and proceeds to process a Reset Indication packet. In either case, the logical the Call Request. channel is placed in the Data Transfer state. Any Data Transfer: Once the logical channel is in Data, Interrupt, Receive Ready (RR), or Receive the Data Transfer state, data, flow control, or reset Not Ready (RNR) packets in the network at the information packets can be transmitted. time of reset are ignored. If a DTE and DCE trans­ The Data packet format includes the packet mit Reset packets simultaneously, a reset collision send and receive sequence numbers. The calling occurs, The net effect is to place the logical channel DTE can transmit up to a predetermined number in the Data Transfer state, ready for Data and In­ of packets before requiring a response. This num­ terrupt packets. ber is referred to as the window. All packet net­ A DTE or a DCE initiates a request for clear­ works must accommodate a lower window edge of ing (disconnecting) a logical channel. Upon confir­ at least eight, permitting at least seven packets to mation by the other device, the logical channel is be outstanding before an acknowledgment is re­ placed in the Ready state. quired. The upper window value (the maximum Error Handling (Packet Level): When an error number of outstanding packets) may not exceed occurs, the DCE transmits Reset, Clear, and Re­ 127 (modulo 128). start indication packets to the DTE. Reset re­ Upon receipt of an authorized packet, the initializes a virtual call or permanent virtual receiving device can either authorize additional transmission by transmitting a Receive Ready

@ 1991 McGraw-Hili, Incorporated. Reproduction Prohibited. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA 22 2720 CCITT Packet Switched Data Networking Networking Standard. Standards XSerle.

Table 7. Packet Types and Their Use in Various Functions " Packet Type Service

Function From DCE to DTE From DTE to DCE Switched Perm. Virtual Virtual Circuit Circuit Can Setup and Cleaning In Incoming Call Call Request X Call Connected Call Accepted X Clear Indication Clear Request X DCE Clear DTE Clear Confirmation Confirmation X Data and Interrupt DCE Data DTE Data X X DCE Interrupt DTE Interrupt X X DCE Interrupt DTE Interrupt Confirmation Confirmation X X Flow Control and Reset DCE RR (Receive DTE RR X X Ready) DTE RNR X X DCE RNR (Receive DTE REJ* X X Not Ready) Reset Request X X Reset Indication DTE Reset DCE Reset Confirmation X X Confirmation Restart Restart Indication Restart Request X X DCE Restart DTE Restart Confirmation Confirmation X X Diagnostics Diagnostics X X

circuit. Reset removes all Data and Interrupt pack­ Implementation Considerations for Users ets on the logical channel and resets the lower win­ X.25 recognizes that DTEs differ in their degree of dow edge to zero. It affects only one logical channel sophistication. A simple DTE may have fixed number (one user). Reset procedures, which apply packet formats and built-in parameter values. only in the Data Transfer state, handle specific er­ while a sophisticated DTE may work with varying ror events, such as local procedure error, remote packet formats and provide variable parameters to procedure error, network congestion, incompatible take advantage of the many network functions and destination, network out of order, etc. Clear affects signaling capabilities offered. only one logical channel number (one user). It At the physical level, the transmission rate clears a session when, for example, the host or host DTE uses to access an X.25 network should be link crashes. In this case, the network initiates governed by the throughput and response time re­ Clear procedures on the terminal link. It effectively quirements of the user. Factors to consider include logs off the terminal user. Clear resets the lower the maximum number of virtual circuits operating window edge to O. Restart affects all users (alliogi­ simultaneously and traffic characteristics of cal channels on a given link) and clears all calls on throughput-critical and response-time-critical vir­ that link. Restart is used when a failed link returns tual circuits. to service; it makes sure that all calls were cleared At the link level, Link Access Procedures when the link went down. Restart also is used (LAPs) and Link Access Procedures Balanced when either side (DTE or DCE) reports error con­ (LAPBs) are defined. A DTE might implement ditions at the packet level; the good side initiates only LAPB. Parameters that are part of the LAPB the restart procedures. procedures can be set to constants for all network In some networks, a diagnostic packet indi­ connections. For example, a timer (T I) may be set cates unusual error conditions. A diagnostic packet according to the slowest required line speed; a con­ from the DCE provides information on events that stant such as three seconds may be used. Also, the are considered unrecoverable at the packet level; maximum permitted number of unacknowledged the information permits the DTE to analyze and I-Frames (information frames) may be determined. initiate recovery by higher levels. No confirmation The network provider must agree to this. All net­ is required on receipt of a diagnostic packet. works support a window of at least seven I-Frames.

FEBRUARY 1991 @ 1991 McGraw-Hili, Incorporated. Reproduction Prohibited. Datapro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 Networking Standarel. 23 X Serle. Standards

The maximum packet size (number of bits) of the link A I and/or data link G I as defined in Recom­ I-Frame must also be established. mendation X.92. The international link should be If the DTE handles certain character sizes capable of supporting full-duplex operation. (octets), the frame level should be capable of ac­ Link-level packet transfer procedures between commodating any number of bits correctly (gener­ STEs consist of the Single Link Procedure (SLP) ate acknowledgment, calculate frame check and the Multilink Procedure (MLP). The SLP is sequence). How such a properly received frame is used for data interchange over a single physical processed depends on the individual system and circuit between two STEs. When multiple physical may include actions such as discarding the frame, circuits are used in parallel. the SLP is used inde­ padding it, clearing the call, or resetting. pendently on each circuit. The MLP is used for data interchange over multiple parallel links. The MLP may be used over a single link when the com­ Connections between Packet Switched municating parties agree to this procedure. Trans­ Data Networks missions are full duplex. SLP is based upon the CCITT Recommendation X.75 describes the pro­ Link Access Procedure Balanced (LAPB) as de­ cedures for the interconnection of packet switched scribed in X.25 and uses the principle and termi­ networks. Many public packet switched data net­ nology of the International Organization for works support X.75 procedures, including AT&T Standardization's (ISO's) High-level Data Link Accunet, TransCanada's Datapac, Graphnet Free­ Control (HDLC). For SLP, either modulo 8 (non­ dom Network II, US Sprint's SprintNet, and extended mode) or the modulo 128 (extended WangPac networks. mode) may be used. The MLP is based on the mul­ X.75 is similar to X.25 in that it specifies pro­ tilink procedures specified by the ISO and per­ cedures' for the physical, link, and packet levels. forms the functions of distributing packets across Signal Terminal Equipment (STE), which acts as a available SLPs. ('I bridge node between networks, implements X. 75 Three channel states are defined: the link procedures. A related standard, CCITT X.121, de­ channel state, which provides transmission in one fines the international numbering plan for public direction; the active channel state, which means data networks. that the incoming or outgoing channel is receiving or transmitting a frame; and the idle channel state, Recommendation X.75 which means that the incoming or outgoing chan­ Recommendation X.75, Terminal and Transmit nel is receiving or transmitting at least 15 contigu­ Call Control Procedures and Data Transfer System ous I bits. on International Circuits Between Packet-Switched Data is transmitted in frames. Each frame Data Networks, provides the rules for transmitting must contain an Opening.Fiag (8 bits), an Address data between different data networks. The basic field (8 bits), a Control field (8 bits), a Frame system structure is made up of communicating ele­ Check Sequence (FCS) field (16 bits), and a Clos­ ments that function independently. These elements ing Flag (8 bits). An Information field of an un­ include the physical circuits, which comprise links specified number of bits, which follows the Control Al or GI (as defined in X.92), and a set of me­ field, is optional. chanical, electrical, functional, and procedural in­ The Control field contains a command or re­ terface characteristics; packet transfer procedures, sponse, and sequence numbers if applicable. Con­ which operate over the physical circuits and pro­ trol field formats may be one of three types: vide for the transport of packets between STEs; numbered information transfer (I format), num­ and packet signaling procedures, which use the bered supervisory functions (S format), and un­ packet data transfer procedures and provide for the numbered control functions (U format). The I exchange of call control information and user traf­ format performs information transfer functions. fic between STEs. Figure 6 shows the basic system The S format performs link supervisory control structure of the signaling and data transfer proce­ dures. The international link is assumed to be data

@ 1991 McGraW-Hili, Incorporated. Reproduction Prohibited FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA' 24 2720 ccm Packet Switched Data Networking Networking Standard. Standards X ...... functions, such as acknowledging I frames, request­ • In/ormation (I) command-transfers sequen­ ing transmission of I frames, and requesting a tem­ tially numbered frames that contain an infor­ porary suspension of transmission of I frames. The mationfield: U format provides additional link control func­ • Receive ready (RR) command and response­ tions. used by the STE to indicate that it is ready to Each I frame is numbered sequentially. The receive an I frame or to acknowledge a previ­ send state variable V(S) represents the sequence ously received I frame: number of the next in-sequence I frame to be trans­ mitted. The value of the V(S) is incremented by • Receive not ready (RNR) command and one with each successive I frame transmission but response-used by the STE to indicate a busy cannot exceed N(R) ofthe last received I or S for­ condition: mat frame by more than the maximum number of • Reject (REI) command and response-used by outstanding frames. The send sequence number the STE to request retransmission of I frames N(S), contained only in I frames, is set equal to the beginning with frame numbered N(R): value of V(S). The receive state variable V(R) rep­ • Set asynchronous balanced mode (SABM) resents the sequence number of the next in­ command-an unnumbered command used to sequence I frame expected to be received. The set the addressed STE in the asynchronous bal­ value of V(R) is incremented by one when an error­ anced mode information transfer phase: all free, in-sequence I frame whose N(S) equals V(R) command/response control fields are one octet is received. Both I frames and S frames contain the (eight bits): receive sequence number N(R). N(R) is the ex­ • Set asynchronous balanced mode extended pected send sequence number of the next received (SABME) command-an unnumbered com­ I frame. When the STE transmits N(R), it indicates mand used to set the addressed STE in the asyn­ that all I frames numbered up to and including chronous balanced mode information transfer N(R)-1 have been received correctly. phase: all numbered command/response control All frames contain the Poll/Final (P/F) bit, fields are two octets: all unnumbered command/ which is referred to as the P bit in command response control fields are one octet: frames and the F bit in response frames. The STE solicits a response from another STE by setting the • Disconnect (DISC) command-terminates the P bit to 1; the answering STE responds by setting previously set mode: DISC indicates that the the F bit to 1. The following commands and re­ STE transmitting DISC is suspending opera­ sponses are supported: tion:

Figure 6. Basic System Stl'llcture ofX.7S Signaling and Data Transfer Procedures r ------I I Higher Level Functions Packet Packet I -Call Control Signaling I-- Transfer ,I I Link -Network Control Procedures Procedures I A1 or G1 -User Traffic I XIV Signaling Terminal (STE) I Interface

Gatewayrrransit Data Switching Exchange X or V

FEBRUARY 1991 @ 1991 McGraw-HIli. Incorporated. ReproductIOn Prohibited. Datapro Information Services Group. Delran NJ 08075 USA Data Networking CCI" Packet Switched 2720 Networking Itandard. X lerle. Standards

• Unnumbered acknowledge (UA) response-used Data transfer procedures apply independently by the STE to acknowledge mode-setting com­ to each logical channel at the X/Y interface. In the mands; data transfer (P4) state, data. interrupt. flow con­ • Disconnected mode (DM) response-reports STE trol, and reset packets may be sent and received by status when it is logically disconnected from the the STE. The data transfer state exists within the link and is in the disconnected phase; and packet-level ready state of a logical channel. Each data packet contains a sequence number called the • Frame reject (FRMR) response-indicates an packet send sequence number peS); only data pack­ error condition that is not recoverable by re­ ets contain the peS). transmission of the frame. The procedures for flow control and reset ap­ ply only in the data transfer state. Flow control ap­ Packet-level signaling procedures relate to the plies to data packets. A window is defined for each transfer of packets at the STE-X/STE-Y (X/Y) in­ direction of transmission at the X/Y interface. The terface. Recommendation X.75 specifies signaling lowest number in the window is called the lower procedures for virtual call setup and clearing, for window edge. The maximum window edge does permanent virtual circuit service, for data and in­ not exceed modulo 8 or 128, is unique to each logi­ terrupt transfer, for flow control, and for reset. cal channel, and is reserved for a period of time. Logical channels are used to complete simul­ For a particular call, two window sizes, one for taneous virtual calls andlor permanent virtual cir­ each direction of transmission, may be selected. cuits. A logical channel group number and a logical The packet receive sequence number peR) (modulo channel number (in the range 0 to 15 inclusive and 8 or 128) conveys information from the receiver oto 255 inclusive, respectively) are assigned to for the transmission of data packets. The peR) be­ each virtual call and permanent virtual circuit. The comes the lower window edge when transmitted logical channel group number and the logical chan­ across the X/Y interface, thereby authorizing addi­ nel number are contained in each packet type, ex­ tional data packets to cross the X/V interface. cept restart packets. Reset procedures are used to reinitialize a The procedures for virtual call setup and single call and apply only in the data transfer state. clearing are used only when a logical channel is in The STE sends a reset request packet that specifies the packet level ready (RL) state. If call setup is the logical channel to indicate a request for reset. possible, and no call or call attempt exists, the logi­ The logical channel is placed in the reset request cal channel is in the ready (PL) state (within the state. The requested STE confirms by sending a RL state). Call setup is initiated when the STE reset confirmation pack~t, which places the logical sends a call request packet across the X/Y inter­ channel in the flow control ready state. face. The call request packet specifies a logical Restart procedures are used to clear all calls channel in the PL state. The logical channel is then simultaneously. When the STE sends a restart re­ placed in the call request state. The called STE in­ quest packet, the X/Y interface for each logical dicates acceptance by the called DTE by sending a channel is placed in the restart request state. In this call connected packet across the X/Y interface. It state all packets, except restart request and restart specifies the same logical channel as that requested confirmation packets, are discarded by the X/Y by the call request packet. The logical channel is interface. An STE confirms by sending a restart then placed in the flow control ready (DL) state confirmation packet, which places all channels in within the data transfer (P4) state. the ready state. A logical channel (in any state) can be cleared Packet formats are based on the general struc­ when the STE sends a clear request packet, which ture of packets as defined in X.25. specifies the logical channel, across the X/Y inter­ face. Upon receipt of a clear request packet, STE-X or STE-Y frees the logical channel and transmits a Trends in Packet Switching clear confirmation packet that specifies the same channel. This places the logical channel in the Support for packet switching has grown among ( ready state within the RL state. Permanent virtual U.S. manufacturers of data processing and data circuits require no call setup or clearing. communications equipment in recent years. For example, nearly all major computer vendors have

CI 1991 McGraw-Hili. Incorporated. Reproductlon Prohibited. FEBRUARY 1991 Datapro Information Service8 Group. Delran NJ 08075 USA 26 2720 COITT Packet Switched Data Networking Networking Standards Standards X Serl•• incorporated X.25 interfaces into their network In 1988. the CCITT also modified the Ad­ architectures. This is part of an overall trend in dress Extension facilities to be consistent with ISO accepting international standards and the increas­ address length. Previously. a provisional 32- ing availability of products conforming to these decimal! 16-octet field had been recommended: this standards. address length was increased to 40 decimals/20 The CCITT published revisions to the X Se­ octets. The ISO also added these address recom­ ries standards in 1984 and in 1989. Since that mendations to Addendum 2 of ISO 8348 time, the ratification and publication of revisions (Connection-Mode Network Service): the CCITT has become a continuous, ongoing process. Since adoption is X.213. the major building blocks for X.25 were laid by 1984, all subsequent changes have been, and will Connectionless Issues: Relationships with ISO continue to be, relatively minor. Some post-1984 Efforts changes have revolved around efforts to make At any layer of the Open Systems Interconnection CCITT X Series standards compatible with those (OSI) Reference Model. two basic forms of opera­ of the International Organization for Standardiza­ tion are possible: connection oriented and connec­ tion (ISO). Currently, discussions on how to pro­ tionless. Connection-oriented service involves a vide greater interoperability between various X.25 connection establishment phase. a data transfer networks is taking place. Major developments in phase, and a connection termination phase. A logi­ packet switching today, however, center not cal connection is set up between end entities prior around X.25, but around the development of new to exchanging data. In a connectionless service. ISDN-related technologies, such as fast packet typical of local area networks. each packet is inde­ switching and frame relay, which support the inte­ pendently routed to the destination. No connection gration of voice, video, and data and much higher establishment activities are required since each transmission speeds. This section discusses post- data unit is independent of the previous or subse­ 1984 changes to X.25 and its relationship to ISDN. quent one. Each transmission mode has a niche where it represents the best approach. For exam­ Major Changes In 1988 Revisions ple, file transfers may benefit from a connection­ In the 1988 revised standards, there were no oriented service, while point-of-sale inquiries may changes at the physical and link levels. At the be best served by a connectionless service. packet level, however, a new facility for redirecting Traditionally, the CCITT has pursued a calls, Call Deflection, was established. In 1984, the connection-oriented philosophy, while ISO has CCITT had made available a new Call Redirection shown interest in connectionless. While the origi­ facility, allowing the network to redirect all calls nal OSI standard, ISO 7498, is connection ori­ destined for a given address. This redirection could ented, ISO saw the need to provide connectionless occur when the destination was out of order or service by issuing an addendum to that protocol, busy, or it could be based on time of day or other ISO 7498/DADl. ISO has issued a standard for criteria. The 1988 facility extended this capability, connection-mode network service (ISO 8348), allowing the destination subscriber to clear incom­ while the CCITT has issued an identical service, ing calls to another party on a call-selective basis. X.213. In regard to X.25 itself, however, ISO has The Clear Request packet contains the Call Deflec­ decided not to pursue the connectionless service, tion information that profiles the desired alternate formerly known as "datagram" service. X.75 is party. also a connection-oriented service; ISO has shown considerable interest in a connectionless internet­ Relationship between CelTf and ISO Efforts working protocol (IP) and has developed the ISO CCITT X.25 packet-level protocol specifies a vir­ 8473 to accommodate it. tual circuit service; the ISO has issued a compati­ ble version of the packet standard, ISO 8208. In Packet Switching in ISDN recent years, CCITT and ISO organizations have The goal of the Integrated Services Digital Net­ worked on standards to carry longer addresses in work (ISDN) is to provide an end-to-end digital the DTE field to facilitate interworking with ISDN path over a set of standardized user interfaces, giv­ (E. 164). ing the user the capability to signal the network

FEBRUARY 1991 @ 1991 McGraW-Hili, Incorporated. Reproduction Prohibited. Datapro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 27 Networking Standard. X Serle. Standards

through an out-of-band channel. (In contrast, in CCITT ISDN Standards X.25, the user signals the network in in-band fash­ ISDN provides a specific protocol that users can ion by issuing packets such as CALL REQUEST, employ to signal the network. Currently. a three­ CALL ACCEPTED, etc.) layer protocol suite is defined. At the Basic Rate. Currently, different types of interfaces to the the Physical Layer manages a 192K bps. full­ telephone network exist for different services. duplex bit stream using time-compression tech­ These interfaces include two-wire switched, two­ niques and time division methods to recover the wire dedicated, four-wire dedicated, DDS, and so two B-channels and one D-channel. The remaining forth. ISDN will provide a small set of interfaces 48K bps stream is used for Physical Layer control that can be used for multiple applications. The information. The defined standards are 1.420 (Ba­ CCITT has defined the following interfaces for sic Rate Interface Definition), 1.430 (Basic Rate ISDN: Interface Layer I). 1.421 (Primary Rate Interface • 2B+D-two 64K bps channels and a 16K bps Definition), and 1.431 (Primary Rate Interface packet/signaling channel (also called the Basic Layer I). Rate Interface). The Data Link Layer is not defined for trans­ parent B-channels used for circuit switched voice • 23B+D-twenty-three 64K bps channels and a or data, but it is defined for the D-channel. For 64K bps packet/signaling channel (also called Narrowband ISDN, the D-channel employs a LAP­ the Primary Rate Interface). o Link Layer protocol, which is a subset of the ISO • 3HO+D-three 384K bps channels and one HDLC Data Link protocol, as specified in CCITT 64K bps packet/signaling channel. Recommendations Q.920 (1.440) and X.921 • Hll-nonchannelized 1.536M bps. (1.441). It provides statistical multiplexing for three channel types: signaling information for the • Hl2-nonchannelized 1.920M bps. management of the B-channels: packet switched • Multislotted-multiples of 64K bps channels service over the D-channel: and optional channels. (up to 1.536M bps) under the customer's con­ used for telemetry of other applications. trol. The Network Layer protocol for the signaling • Broadband-high data rates, based on an ap­ channel is specified in CCITT's Q.930 (1.450) and proach called synchronous optical network Q.931 (1.451) specifications. It provides the mech­ (SONET), building on multiplex of 51.84M bps. anism for establishing and terminating connections SONET standards negotiations began in 1986. on the B-channels and other network control func­ The CCITT approved phase I of the standards tions. For the packet switched service over the D­ in 1988 and phase II in 1989. This architecture channel, the Network Layer protocol is X.25. has been called Broadband ISDN, in contrast to CCITT will define Layer 3 protocols for the op­ the other interfaces that have been considered tional channels in the future, or they will be speci­ part of Narrowband ISDN. fied as national options. A technique is required for specifying With the exception of Broadband ISDN, all of the whether user-to-network signaling. user packet above interfaces could be carried on unloaded cop­ data, or user telemetry data is being sent over the per loops. Using fiber has also been considered, as D-channel. This technique involves the use of a it would make the local loop more robust. Out-of­ service access point identifier (SAPI). band signaling makes possible a new class of ser­ Each layer in the OSI Reference Model com­ vices. In addition, the 16K bps D-channel will be municates with the layers above and below it connected directly to the BOC's packet switched across an interface. The interface is through one or network, providing the subscriber with the data more service access points (SAPs). SAPs have a multiplexing advantages packet switching offers. A number of uses, including subaddressing for inter­ major effort is under way in Europe to bring the networking situations, Transport Layer applica­ system to market. In the United States, several tri­ tions, and for user data packet service over an als have been undertaken, and limited ISDN ser­ ISDN D-channel. vice is already available. Considering the applications to the Transport Layer, one should note that two general types of

@ 1991 McGraw-Hili, Incorporated. Reproduction Prohibited. FEBRUARY 1991 Datepro Information Services Group. Delran NJ 08075 USA 28 2720 oem Packet SwHcbed Data Networking Networking Standa"'" Standards X Serle.

Figure 7. SAPl Action for a Basic Rate D-Channe/ -_/

Signaling Packet Telemetry ------_.Layer 4

Layer 3 Layer 3 Network Layer 3 Layer 3 Layer 3 s Entity P Entity 81 Entity 82 Entity t Entity r------,=-ai'~r_ -- --.. (SAP1=O) (SAP1=16) Protocol

Layer 3 ------r------y r- Layer 2 Layer 2 Layer 2 Data link 81 Entity 82 Entity LAPD Entity 64K bps 64K bps 64K bps ~-----~y~~-----~Protocols

Layer 2 , t Layer 1 ISDN Physical Entity Physical Layer Protocols 192K bps ~------~

------_.Layer 1 Physical Media (ISDN link) ~------.. addressing in a communications architecture are traffic type and the Data Link Layer services di­ available. Each host on the network must have a rected to Layer 3. For example, the SAPI value of 0 network address, allowing the network to deliver directs the frames to Layer 3 for call-control proce­ data to the proper computer. Each process within a dures; a SAPI value of 16 indicates a packet com­ host must have an address that is unique within the munication procedure. See Figure 7. host; this allows the Transport Layer to deliver data to the proper process. These process addresses Fast Packet Switching are identified using SAPs. A similar approach is Faster switches are required before packet­ followed for ISDN. switching service is more widely used. Current LAP-D, the data link standard for ISDN, packet switches can provide a throughput of up to specifies the link access protocol used on the D­ 40,000 packets per second. The fast packet­ channel. LAP-D is based on LAP-B, which is based switching technique has drawn considerable inter­ on HDLC. LAP-D must deal with two levels of est, since vendors of fast packet products promise multiplexing. First, at a subscriber location, up to 800,000 packets per second throughput. Ac­ multiple-user devices may be sharing the same cording to most sources, carriers throughout the physical interface. Second, each user device may world view fast packet technology as the technique support mUltiple types of traffic, including packet of choice for future networks. switched data and signaling. To accomplish this In the fast packet environment, voice, data type of multiplexing, LAP-D employs a two-part signaling, video, mass-memory transfers, and high­ address consisting of a terminal endpoint identifier speed LAN interconnections are combined and (TEl) and a SAPI. Typically, each user terminal is channeled through a common physical network. given a distinguishing TEL The SAPI identifies the This technology provides continuously adaptive

FEBRUARY 1991 ~ 1991 McGraw-Hili, Incorporated. Reproduction Prohibited. Datepro Information Services Group. Delran NJ 08075 USA Data Networking CCITT Packet Switched 2720 29 Networking Standard. X Serle. Standards

bandwidth management, allowing each user to ac­ intermediate network nodes are assumed caught quire the requisite amount of transport capacity on and corrected by higher layer protocols. Thus. in­ an as-needed basis. termediate nodes simply forward packets (called Data packets produced by the fast packet frames) without processing the datastream. In ad­ technique are like samples produced through pulse dition, frames must be received in the order in code modulation (PCM). Since X.25 data packets which they were sent. unlike some X.25 networks. can contain data only, traditional packet switching which involves considerably less machine process­ is not compatible with fast packet technology. Pro­ ing at the opposite end of transmission. These effi­ prietary solutions are available, however, from ciencies result in extremely higher throughput Stratacom and other vendors, for integrating both speeds-up to 2M bps. types of systems. Most commercial frame relay products are based on ISDN recommendations contained in Frame Relay CCITT 1.122, entitled "Framework for Providing Frame relay is an emerging, standards-based ad­ Additional Packet Mode Bearer Services," and/or dressing technique that has great potential in fast in ANSI Tl S 1/88-2242, "Frame Relay Bearer packet swithing and internetworking of remote Service-Architecture and Description." Cur­ LANs. Frame relay is analogous to how X.25 re­ rently, only one packet service has been specified lates to conventional packet-switching backbone in ISDN standards: "Support of Packet Mode Ter­ networks. X.25 is a "network access method," or minal Equipment by an ISDN," CCITT 1.462 user-to-network interface. Frame relay provides (X.31). More work is being accomplished during user access to a higher speed network that is based the 1989-1992 Study Period, however, to specify on a transmission technology other than X.25. additional packet switched services. Vendors have recently developed frame relay Vendors from several different networking interfaces for conventional packet and circuit disciplines are trying to establish themselves as switched networks. LAN internetwork and WAN frame relay proponents. These include T -carrier vendors are also teaming up to develop frame relay nodal processor vendors; LAN internetwork ven­ capabilities for remote bridges and routers, offer­ dors (bridges and routers); traditional packet ing integrated LAN/WAN solutions to customers. switched network providers, such as Hughes Net­ Frame relay is based on the CCITT Layer 2 work Systems, Netrix, Telematics, and US Sprint; protocol developed for ISDN, Link Access Proto­ and communications carriers. Many have an­ col D (LAPD). Unlike conventional X.25 packet nounced produ~ts or services that will be available switching, frame relay uses variable packet lengths in 1991_ Stratacom is one vendor that has been and performs error checking only at the remote instrumental in developing frame relay concepts end 'of transmission. Any errors occurring between (along with fast packet technology) .•

(

@ 1991 McGraw-Hili. Incorporated. Reproduction Prohibited. FEBRUARY 1991 Datapro Information Services Group. Delran NJ 08075 USA