Integrated Services Digital Network (ISDN) is a set of communication standards for simultaneous digital transmission of voice, video, data, and other network services over the digitalised circuits of the public switched telephone network.[1] Work on the standard began in 1980 at Bell Labs and was formally standardized in 1988 in the CCITT "Red Book".[2] By the time the standard was released, newer networking systems with much greater speeds were available, and ISDN saw relatively little uptake in the wider market. One estimate suggests ISDN use peaked at a worldwide total of 25 million subscribers at a time when 1.3 billion analog lines were in use.[3] ISDN has largely been replaced with digital subscriber line (DSL) systems of much higher performance.
Prior to ISDN, the telephone system consisted of digital links like T1/E1 on the long-distance lines between telephone company offices and analog signals on copper telephone wires to the customers, the "last mile". At the time, the network was viewed as a way to transport voice, with some special services available for data using additional equipment like modems or by providing a T1 on the customer's location. What became ISDN started as an effort to digitize the last mile, originally under the name "Public Switched Digital Capacity" (PSDC).[3] This would allow call routing to be completed in an all-digital system, while also offering a separate data line. The Basic Rate Interface, or BRI, is the standard last-mile connection in the ISDN system, offering two 64 kbit/s "bearer" lines and a single 16 kbit/s "data" channel for commands and data.
Although ISDN was successful in a few countries such as Germany, on a global scale the system was largely ignored and garnered the industry nickname "innovation subscribers didn't need."[4] It found a use for a time for small-office digital connection, using the voice lines for data at 64 kbit/s, sometimes "bonded" to 128 kbit/s, but the introduction of 56 kbit/s modems undercut its value in many roles. It also found use in videoconference systems, where the direct end-to-end connection was desirable. The H.320 standard was designed around its 64 kbit/s data rate. The underlying ISDN concepts found wider use as a replacement for the T1/E1 lines it was originally intended to extend, roughly doubling the performance of those lines.
Since its introduction in 1881, the twisted pair copper line has been installed for telephone use worldwide, with well over a billion individual connections installed by the year 2000. Over the first half of the 20th century, the connection of these lines to form calls was increasingly automated, culminating in the crossbar switches that had largely replaced earlier concepts by the 1950s.[3]
As telephone use surged in the post-WWII era, the problem of connecting the massive number of lines became an area of significant study. Bell Labs' seminal work on digital encoding of voice led to the use of 64 kbit/s as a standard for voice lines (or 56 kbit/s in some systems). In 1962, Robert Aaron of Bell introduced the T1 system, which carried 1.544 Mbit/s of data on a pair of twisted pair lines over a distance of about one mile. This was used in the Bell network to carry traffic between local switch offices, with 24 voice lines at 64 kbit/s and a separate 8 kbit/s line for signaling commands like connecting or hanging up a call. This could be extended over long distances using repeaters in the lines. T1 used a very simple encoding scheme, alternate mark inversion (AMI), which reached only a few percent of the theoretical capacity of the line but was appropriate for 1960s electronics.[4]
By the late 1970s, T1 lines and their faster counterparts, along with all-digital switching systems, had replaced the earlier analog systems for most of the western world, leaving only the customer's equipment and their local end office using analog systems. Digitizing this "last mile" was increasingly seen as the next problem that needed to be solved. However, these connections now represented over 99% of the total telephony network, as the upstream links had increasingly been aggregated into a smaller number of much higher performance systems, especially after the introduction of fiber optic lines. If the system was to become all-digital, a new standard would be needed that was appropriate for the existing customer lines, which might be miles long and of widely varying quality.[4]
Around 1978, Ralph Wyndrum, Barry Bossick and Joe Lechleider of Bell Labs began one such effort to develop a last-mile solution. They studied a number of derivatives of the T1's AMI concept and concluded that a customer-side line could reliably carry about 160 kbit/s of data over a distance of 4 to 5 miles (6.4 to 8.0 km). That would be enough to carry two voice-quality lines at 64 kbit/s as well as a separate 16 kbit/s line for data. At the time, modems were normally 300 bit/s and 1200 bit/s would not become common until the early 1980s and the 2400 bit/s standard would not be completed until 1984. In this market, 16 kbit/s represented a significant advance in performance in addition to being a separate channel that coexists with voice channels.[4]
A key problem was that the customer might only have a single twisted pair line to the location of the handset, so the solution used in T1 with separate upstream and downstream connections was not universally available. With analog connections, the solution was to use echo cancellation, but at the much higher bandwidth of the new concept, this would not be so simple. A debate broke out between teams worldwide about the best solution to this problem; some promoted newer versions of echo cancellation, while others preferred the "ping pong" concept where the direction of data would rapidly switch the line from send to receive at such a high rate it would not be noticeable to the user. John Cioffi had recently demonstrated echo cancellation would work at these speeds, and further suggested that they should consider moving directly to 1.5 Mbit/s performance using this concept. The suggestion was literally laughed off the table (His boss told him to "sit down and shut up"[4]) but the echo cancellation concept that was taken up by Joe Lechleider eventually came to win the debate.[4]
Meanwhile, the debate over the encoding scheme itself was also ongoing. As the new standard was to be international, this was even more contentious as several regional digital standards had emerged in the 1960s and 70s and merging them was not going to be easy. To further confuse issues, in 1984 the Bell System was broken up and the US center for development moved to the American National Standards Institute (ANSI) T1D1.3 committee. Thomas Starr of the newly formed Ameritech led this effort and eventually convinced the ANSI group to select the 2B1Q standard proposed by Peter Adams of British Telecom. This standard used an 80 kHz base frequency and encoded two bits per baud to produce the 160 kbit/s base rate. Ultimately Japan selected a different standard, and Germany selected one with three levels instead of four, but all of these could interchange with the ANSI standard.[5]
From an economic perspective, the European Commission sought to liberalize and regulate ISDN across the European Economic Community.[6] The Council of the European Communities adopted Council Recommendation 86/659/EEC Archived 2022-10-18 at the Wayback Machine in December 1986 for its coordinated introduction within the framework of CEPT. ETSI (the European Telecommunications Standards Institute) was created by CEPT in 1988 and would develop the framework.
With digital-quality voice made possible by ISDN, offering two separate lines and continuous data connectivity, there was an initial global expectation of high customer demand for such systems in both the home and office environments. This expectation was met with varying degrees of success across different regions.
In the United States, many changes in the market led to the introduction of ISDN being tepid. During the lengthy standardization process, new concepts rendered the system largely superfluous. In the office, multi-line digital switches like the Meridian Norstar took over telephone lines while local area networks like Ethernet provided performance around 10 Mbit/s which had become the baseline for inter-computer connections in offices. ISDN offered no real advantages in the voice role and was far from competitive in data. Additionally, modems had continued improving, introducing 9600 bit/s systems in the late 1980s and 14.4 kbit/s in 1991, which significantly eroded ISDN's value proposition for the home customer.[5]
Conversely, in Europe, ISDN found fertile ground for deployment, driven by regulatory support, infrastructural needs, and the absence of comparable high-speed communication technologies at the time. The technology was widely embraced for its ability to digitalize the "last mile" of telecommunications, significantly enhancing the quality and efficiency of voice, data, and video transmission over traditional analog systems.
Meanwhile, Lechleider had proposed using ISDN's echo cancellation and 2B1Q encoding on existing T1 connections so that the distance between repeaters could be doubled to about 2 miles (3.2 km). Another standards war broke out, but in 1991 Lechleider's 1.6 Mbit/s "High-Speed Digital Subscriber Line" eventually won this process as well, after Starr drove it through the ANSI T1E1.4 group. A similar standard emerged in Europe to replace their E1 lines, increasing the sampling range from 80 to 100 kHz to achieve 2.048 Mbit/s.[7] By the mid-1990s, these Primary Rate Interface (PRI) lines had largely replaced T1 and E1 between telephone company offices.
Lechleider also believed this higher-speed standard would be much more attractive to customers than ISDN had proven. Unfortunately, at these speeds, the systems suffered from a type of crosstalk known as "NEXT", for "near-end crosstalk". This made longer connections on customer lines difficult. Lechleider noted that NEXT only occurred when similar frequencies were being used, and could be diminished if one of the directions used a different carrier rate, but doing so would reduce the potential bandwidth of that channel. Lechleider suggested that most consumer use would be asymmetric anyway, and that providing a high-speed channel towards the user and a lower speed return would be suitable for many uses.[7]
This work in the early 1990s eventually led to the ADSL concept, which emerged in 1995. An early supporter of the concept was Alcatel, who jumped on ADSL while many other companies were still devoted to ISDN. Krish Prabu stated that "Alcatel will have to invest one billion dollars in ADSL before it makes a profit, but it is worth it." They introduced the first DSL Access Multiplexers (DSLAM), the large multi-modem systems used at the telephony offices, and later introduced customer ADSL modems under the Thomson brand. Alcatel remained the primary vendor of ADSL systems for well over a decade.[8]
ADSL quickly replaced ISDN as the customer-facing solution for last-mile connectivity. ISDN has largely disappeared on the customer side, remaining in use only in niche roles like dedicated teleconferencing systems and similar legacy systems.
Integrated services refers to ISDN's ability to deliver at minimum two simultaneous connections, in any combination of data, voice, video, and fax, over a single line. Multiple devices can be attached to the line, and used as needed. That means an ISDN line can take care of what were expected to be most people's complete communications needs (apart from broadband Internet access and entertainment television) at a much higher transmission rate, without forcing the purchase of multiple analog phone lines. It also refers to integrated switching and transmission[9] in that telephone switching and carrier wave transmission are integrated rather than separate as in earlier technology.
In ISDN, there are two types of channels, B (for "bearer") and D (for "data"). B channels are used for data (which may include voice), and D channels are intended for signaling and control (but can also be used for data).
There are two ISDN implementations. Basic Rate Interface (BRI), also called basic rate access (BRA) — consists of two B channels, each with bandwidth of 64 kbit/s, and one D channel with a bandwidth of 16 kbit/s. Together these three channels can be designated as 2B+D. Primary Rate Interface (PRI), also called primary rate access (PRA) in Europe — contains a greater number of B channels and a D channel with a bandwidth of 64 kbit/s. The number of B channels for PRI varies according to the nation: in North America and Japan it is 23B+1D, with an aggregate bit rate of 1.544 Mbit/s (T1); in Europe, India and Australia it is 30B+2D, with an aggregate bit rate of 2.048 Mbit/s (E1). Broadband Integrated Services Digital Network (BISDN) is another ISDN implementation and it is able to manage different types of services at the same time. It is primarily used within network backbones and employs ATM.
Another alternative ISDN configuration can be used in which the B channels of an ISDN BRI line are bonded to provide a total duplex bandwidth of 128 kbit/s. This precludes use of the line for voice calls while the internet connection is in use. The B channels of several BRIs can be bonded, a typical use is a 384K videoconferencing channel.
Using bipolar with eight-zero substitution encoding technique, call data is transmitted over the data (B) channels, with the signaling (D) channels used for call setup and management. Once a call is set up, there is a simple 64 kbit/s synchronous bidirectional data channel (actually implemented as two simplex channels, one in each direction) between the end parties, lasting until the call is terminated. There can be as many calls as there are bearer channels, to the same or different end-points. Bearer channels may also be multiplexed into what may be considered single, higher-bandwidth channels via a process called B channel BONDING, or via use of Multi-Link PPP "bundling" or by using an H0, H11, or H12 channel on a PRI.
The D channel can also be used for sending and receiving X.25 data packets, and connection to X.25 packet network, this is specified in X.31. In practice, X.31 was only commercially implemented in the UK, France, Japan and Germany.
A set of reference points are defined in the ISDN standard to refer to certain points between the telco and the end-user equipment.
Most NT-1 devices can perform the functions of the NT2 as well, and so the S and T reference points are generally collapsed into the S/T reference point.
In North America, the NT1 device is considered customer premises equipment (CPE) and must be maintained by the customer, thus, the U interface is provided to the customer. In other locations, the NT1 device is maintained by the telco, and the S/T interface is provided to the customer. In India, service providers provide U interface and an NT1 may be supplied by Service provider as part of service offering.
The entry level interface to ISDN is the Basic Rate Interface (BRI), a 128 kbit/s service delivered over a pair of standard telephone copper wires.[10] The 144 kbit/s overall payload rate is divided into two 64 kbit/s bearer channels ('B' channels) and one 16 kbit/s signaling channel ('D' channel or data channel). This is sometimes referred to as 2B+D.[11]
The interface specifies the following network interfaces:
BRI-ISDN is very popular in Europe but is much less common in North America. It is also common in Japan — where it is known as INS64.[12][13]
The other ISDN access available is the Primary Rate Interface (PRI), which is carried over T-carrier (T1) with 24 time slots (channels) in North America, and over E-carrier (E1) with 32 channels in most other countries. Each channel provides transmission at a 64 kbit/s data rate.
With the E1 carrier, the available channels are divided into 30 bearer (B) channels, one data (D) channel, and one timing and alarm channel. This scheme is often referred to as 30B+2D.[14]
In North America, PRI service is delivered via T1 carriers with only one data channel, often referred to as 23B+D, and a total data rate of 1544 kbit/s. Non-Facility Associated Signalling (NFAS) allows two or more PRI circuits to be controlled by a single D channel, which is sometimes called 23B+D + n*24B. D-channel backup allows for a second D channel in case the primary fails. NFAS is commonly used on a Digital Signal 3 (DS3/T3).
PRI-ISDN is popular throughout the world, especially for connecting private branch exchanges to the public switched telephone network (PSTN).
Even though many network professionals use the term ISDN to refer to the lower-bandwidth BRI circuit, in North America BRI is relatively uncommon whilst PRI circuits serving PBXs are commonplace.
The bearer channel (B) is a standard 64 kbit/s voice channel of 8 bits sampled at 8 kHz with G.711 encoding. B-channels can also be used to carry data, since they are nothing more than digital channels.
Each one of these channels is known as a DS0.
Most B channels