ISDN

Integrated Services Digital Network (ISDN) started becoming popular in the U.S. in the mid-1990s with the growing popularity of the Internet and with demand for cheap, high-speed circuits.

Although the telephone network began switching from analog to digital circuits back in the 1960s, it wasn't until 1984 that the International Telecommunication Union Telecommunication Standardization Sector (ITU-T) specified a set of standards. Although some areas of the world cleanly implemented the standards, lack of a single government carrier in the U.S. led to multiple, incompatible implementations of the standards. U.S. carriers ratified the network termination (NI-1) standard in the early 1990s. It was supposed to allow for increased compatibility between the systems, but some providers refused to implement it. Cost and backward compatibility with their previous implementation were often cited.

Because consumers wanted access to more content than the average 33.6Kbps modem could easily provide, telcos were under pressure to roll out ISDN to residential areas. Those telcos that refused were occasionally subject to lawsuits and sometimes they lost.

ISDN comes in two forms: Basic Rate Interface (BRI) and Primary Rate Interface (PRI). Use and functionality of the two types differ around the world. All ISDN connections use the 64Kbps circuit; what differs is how many there are and what the purpose of each is. ISDN uses pre-established switched circuits.

BRI

BRI connections use a total of three 64Kbps channels, called DS-0s. Two of these channels may be used for data or voice and are called bearer or B channels. Because there are two channels, it is possible to use one for voice and the other for data at the same time. B channels can be either up or down, depending on the need. If a call is being made, they are active. An inactivity timer or application sensor can shut down a connection when the connection is no longer needed.

The third circuit is often called the delta or D channel. Because this circuit handles signaling information with the telco, it is always on. The telco signaling uses only 48Kbps, leaving 16Kbps free. The ISDN devices themselves can use the remaining bandwidth to provide authentication and other connection information. This setup allows an organization to implement security while not incurring costs by bringing up B channels unnecessarily. Figure 7.1 shows the B and D channels.

Figure 7.1. BRI channels.

In the U.S., ISDN originally started on a pay-per-minute basis and stayed that way in many places until the late 1990s when most telcos began offering ISDN on a monthly subscription rate. Many other areas of the world still use pay-per-minute per B channel in use. ISDN and Cisco routers have several capabilities for reducing the amount of money spent on idle ISDN connections.

64Kbps is the optimum bandwidth on a circuit, but this value might be lower, depending on the telco configuration. Alternate Mark Inversion (AMI) encoding is common throughout the industry and often results in inconsistent connectivity. Often, setting the ISDN channel bandwidth to 56Kbps results in the connection coming right up.

PRI

A PRI is a larger bundle of 64Kbps channels that has two different worldwide implementations. The T1 exists mostly in the Western hemisphere and consists of 24 DS-0s, whereas the E1 primarily appears in the Eastern hemisphere and consists of 31 DS-0s. There isn't any reason why you can't order an E1 in the U.S. or a T1 in Europe because it all depends on what the telco has implemented and is willing to configure. The only differences are that one supports more channels than the other and you must correctly set the framing of the bits.

There isn't a distinct D channel, although there can be, because signaling can happen in a couple of different ways. First, a channel can be set aside and used for signaling for the other 23 channels. This setup is sometimes referred to as the 23+1 configuration for 23 B channels and one D channel. The other way involves taking a small amount of bandwidth from each of the DS-0s and using it for signaling. This setup results in slightly less bandwidth per channel, but all channels can be used.

ISDN Usage

The channels are separated into 64Kbps channels because that's what is needed for a voice connection using 8 bits being sampled at 8,000 times per second. Voice connections need to keep the channels separated, but if some are used for data connectivity, it is more efficient to bundle them together.

Routers and other devices can logically tie circuits together so they appear to be all one connection. A T1 has 24 64Kbps channels coming in, but bundled together, they look like a single 1.536Mbps circuit. When the 8Kbps of T1 overhead is accounted for, the interface can report 1.544Mbps. The overhead is used for managing the circuit, as explained later.

It's common for an organization to split the tasks that the DS-0s are performing. Some might be voice circuits and others might be bundled together for data transfer. The voice circuits go from the channel service unit/data service unit (CSU/DSU) to a private branch exchange (PBX), while those handling data are sent to the router. If the router has a built-in CSU, it can route the voice channels itself, as in Voice over IP (VoIP) implementations.