304 VOICE OVER ATM AND MPLS
circuit across an ATM network. The incoming bits from the DS1 circuit are simply placed sequentially into the payload of the AAL 1, without regard to framing, using the
unstructured data transfer protocol see Section 3.7.1.
12.4 THE ATM TRUNKING USING AAL 2 FOR NARROWBAND SERVICES
SPECIFICATION
The ATM trunking using AAL 2 for narrowband services specification was designed to interconnect two distant public or private telephone networks over an ATM network. It
can be used, for instance, to interconnect a distant PBX and a central office over an ATM network, such as PBX A and the central office in Figure 12.10. It can be also used to
connect two distant PBXs over an ATM network, such as PBX B and C in Figure 12.10. A PBX or a central office is connected to an IWF over a T1E1 link. This specification
is used in cellular telephony to transport multiple voice calls.
The protocol stack of an IWF that supports the ATM trunking using AAL 2 for narrow- band service
specification is shown in Figure 12.11. As can be seen, an IWF can transport
IWF T1E1
IWF ATM
network Central
office PBX A
PBX B IWF
IWF PBX C
T1E1 T1E1
T1E1
Figure 12.10 ATM trunking using AAL 2 for narrowband services.
Circuit mode data services
Nx64 Kbps User traffic
Voiceband services Inband
signaling User traffic
PCM voice
Compressed voice
Facsimile demodulation
AAL 2 SSCS for trunking SEG-SSCS
Frame mode data services
User traffic IWF-IWF
CPS AAL 5
ATM layer
Figure 12.11 The protocol stack of an IWF.
THE ATM TRUNKING USING AAL 2 305
PCM voice i.e., 64-Kbps voice, compressed voice, facsimile, in-band signaling, and cir- cuit mode data i.e., fractional T1E1. In addition, it can transport frame mode data and
signaling between two IWFs. The ATM trunking using AAL 2 for narrowband service specification uses the services
of two AAL 2 service-specific convergence sublayers developed explicitly for voice trunk- ing: the AAL 2 SSCS for trunking and the segmentation and reassembly service-specific
convergence sublayer for AAL 2 SEG-SSCS . The AAL SSCS for trunking is described
in detail below in Section 12.5. The purpose of this sublayer is to convey telephone voice calls, voiceband data, such as facsimile and data transmitted over a modem, and fractional
T1E1 data. SEG-SSCS for AAL 2 is described in detail in Section 12.6. Using this sub- layer, it is possible to transport a packet with a size bigger than the maximum length of
45 bytes permitted in the payload of the CPS packet.
12.4.1 Switched and Non-Switched Trunking
The ATM trunking using AAL 2 for narrowband services specification covers both switched trunking
and non-switched trunking. Switched trunking involves analysis of the signaling that accompanies an incoming narrowband call and routing of the user data to an AAL 2
channel that runs over an ATM connection. Similar analysis is required for the incoming calls from an ATM network. There is no permanent correspondence between a TDM time
slot to an AAL 2 channel and ATM connection. A new call on the same TDM time slot can be switched to a different AAL 2 connection and ATM connection.
In non-switched trunking, the information stream of a narrowband channel is always carried on the same AAL 2 channel within the same ATM connection. That is, there is
a permanent correspondence between a narrowband call and an AAL 2 channel over a specific ATM connection. Non-switched trunking involves no termination of signaling
and no routing of narrowband calls in the IWFs.
12.4.2 IWF Functionality for Switched Trunking
The narrowband signaling associated with the individual 64-Kbps channels, whether CAS or CCS, is extracted and forwarded to the signaling termination and call handling function
of the IWF. The destination egress IWF is determined based on the calling party address. In CAS the calling party address is signaled using the dual-tone multi-frequency DTMF
system. In CCS, it is signaled in the setup message. CCS information is forwarded to the destination IWF using the SAAL, or using AAL 2 SEG-SSCS SSADT described below
in Section 12.6. CAS bits are forwarded to the destination IWF using the special CAS bit packet defined in Section 12.5.
The signaling termination and call handling function also communicates with the CID management function of AAL 2 to assign a CID value for a new call. This CID value
represents an AAL 2 channel over some ATM connection. If there is not enough bandwidth over the existing ATM connections to the destination IWF, a new ATM connection is
created using the ATM signaling procedures.
12.4.3 IWF Functionality for Non-switched Trunking
The IWF does not interpret, or respond, or process incoming signals. Instead, signaling are passed transparently between the narrowband side and the ATM side. Time slots on
306 VOICE OVER ATM AND MPLS
the incoming T1E1 link are permanently mapped to AAL 2 channels. CCS information is carried either over SEG-SSCS SSTED or AAL 5. ABCD CAS bits and DTMF
information are mapped into the same AAL 2 channel as the user information using the special CAS bit packet defined in the following section.
12.5 THE AAL 2 SERVICE-SPECIFIC CONVERGENCE SUBLAYER SSCS
FOR TRUNKING
The purpose of this convergence sublayer is to convey voice calls, voiceband data, such as facsimile and data transmitted over a modem, and fractional T1E1 data.
The reference model of the AAL 2 SSCS for trunking is shown in Figure 12.12. This model depicts a single connection. On either side of the connection, there is a
transmitting and a receiving SSCS. For each transmitting and receiving SSCS, there is a signal processing device, which passes and receives information to and from the SSCS.
This device is called a User and is indicated by a capital U to distinguish it from a user, i.e., a customer, who generates the traffic carried over this AAL 2 connection.
At the transmitting side, the User provides a number of functions, such as encoding of voice signals, extraction of dialed digits from multi-frequency tones, and extraction of
the ABCD CAS bits. On the receiving side, it decodes voice, generates multi-frequency tones from the received dialed digits, and regenerates the ABCD CAS bits.
The SSCS runs on top of CPS. At the transmitting side, SSCS uses a number of different CPS-packet formats to transport the data received from its User. The CPS-packets are
passed on to CPS and are eventually delivered to the receiving SSCS, from where the data is extracted and passed on to its User.
12.5.1 User Functions
The following are some of the functions provided by the User: Audio
At the transmitter’s side, it encodes audio samples using one of several audio algorithms. The transmitting User also detects silence periods and sends silence insertion descriptors.
At the receiver’s side, it decodes audio bits into a sequence of audio samples, including comfort noise generation as directed by silence insertion descriptors.
Various encoding algorithms can be used. Each algorithm creates encodings, which are grouped together into a packet referred to as the encoding data unit EDU. Bigger
packets can be formed by concatenating several EDUs. The User passes the EDUs to the SSCS, which transmits them to the destination SSCS, which in turn forwards them
User SSCS
transmitter User
SSCS receiver
SSCS receiver
AAL 2 connection
User SSCS
transmitter User
Figure 12.12 The reference model for the AAL 2 SSCS for trunking.
THE AAL 2 SERVICE-SPECIFIC CONVERGENCE SUBLAYER 307
to the destination User, which is responsible for decoding them into a sequence of audio samples. The following are some of the ITU-T audio algorithms.
• G.711 pulse code modulation PCM
: Produces one 8-bit value every 125 µsec, repre- senting the sign and amplitude of an audio sample. Two encoding laws – the A-law and
the µ-law – can be used. It normally transmits at 64 Kbps, but it can also transmit at 56 Kbps and 48 Kbps. The G.711 output is accumulated over 1 msec to form an EDU
of 8 encoded values.
• G.722 sub-band adaptive pulse code modulation SB-ADPCM
: Produces one 8-bit value every 125 µsec, and represents audio samples with higher fidelity than G.711 PCM. The
EDU consists of eight encoded values i.e. values collected over 1 msec. •
G.723.1 : Operates at either 5.3 or 6.4 Kbps. Both rates are a mandatory part of the
encoder and decoder. Every 30 ms, it emits either 160 or 192 bits, respectively; this characterizes a voice sample. It is possible to switch between the two rates at any 30
msec boundary. •
G.726 adaptive pulse code modulation ADPCM : Supports bit rates of 40, 32, 24, and
16 Kbps. Every 125 µsec, the encoding produces 5, 4, 3, or 2 bits, respectively. •
G.722 embedded adaptive pulse code modulation EADPCM : This is a family of vari-
able bit rate coding algorithms with the capability of bit dropping outside the encoder and decoder blocks. It produces code words which contain enhancement bits and core
bits . The enhancement bits can be discarded during network congestion. The number
of code bits must remain the same to avoid mistracking of the adaptation state between transmitter and receiver. Algorithms of the G.727 family are referred to by the pair x,
y, where x is the number of core plus enhancement bits and y is the number of core bits. Recommendation G.727 provides coding rates of 40, 32, 24, and 16 Kbps, with
core rates of 16, 24, and 32 Kbps. This corresponds to the following x, y pairs: 5, 2, 4, 2, 3, 2, 2, 2, 5, 3, 4, 3, 3, 3, 5, 4, and 4, 4. The data unit format requires
that G.727 outputs be accumulated over an interval of 1 msec to yield a sequence of eight encoded values.
• G.728 low delay code excited linear prediction LD-CELP
: This coder produces a group of four codewords every 2.5 msec. Each group of codewords is referred to as an
adaptation cycle or frame.
• G.729
: Runs at 8 Kbps. Every 10 msec, it emits 80 bits that encode a voice frame. Multi-frequency tones and CAS bits
At the transmitter’s side, it detects and extracts dialed digits codes from multi-frequency tones, such as DTMF. It also extracts the ABCD CAS bits. At the receiver’s side, it
regenerates the multi-frequency tones from the received dialed digit codes and regenerates the ABCD CAS bits.
Facsimile At the transmitter’s side, when it detects a facsimile, it demodulates the facsimile signal
and sends the demodulated original image data and associated control signals to the transmitting SSCS. At the receiver’s side, it receives the image data and control signals
from the receiving SSCS, it remodulates them into voiceband for transmission to the peer facsimile terminal. This demodulationremodulation procedure provides a higher-fidelity
transfer.