Voice Interface Configuration

Voice gateways bridge the gap between the VoIP world and the traditional telephony world (for example, a PBX, the PSTN, or an analog phone). Cisco voice gateways connect to traditional telephony devices via voice ports. This chapter introduces basic configuration of analog and digital voice ports, and demonstrates how to fine-tune voice ports with port-specific configurations.

Upon completing this chapter, you will be able to configure voice interfaces on Cisco voice-enabled equipment for connection to traditional, nonpacketized telephony equipment.


Configuring Voice Ports

Connecting voice devices to a network infrastructure requires an in-depth understanding of signaling and electrical characteristics that are specific to each type of interface. Improperly matched electrical components can cause echo and make a connection unusable. As another consideration, configuring devices for international implementation requires knowledge of country-specific settings. This section provides voice port configuration parameters for signaling and country-specific settings.

Before delving into the specific syntax of configuring these voice ports, this section begins by considering several examples of voice applications. The applications discussed help illustrate the function of the voice ports, whose configuration is addressed at the end of this section.


Voice Applications

Different types of applications require specific types of ports. In many instances, the type of port is dependent on the voice device connected to the network. Different types of voice applications include the following:

• Local calls

• On-net calls

• Off-net calls

• Private line, automatic ringdown (PLAR) calls

• PBX-to-PBX calls

• CallManager-to-CallManager calls

• On-net to off-net calls

Local Calls
Local calls, as illustrated in Figure 3-1, occur between two telephones connected to one Cisco voice-enabled router. This type of call is handled entirely by the router and does not travel over an external network. Both telephones are directly connected to Foreign Exchange Station (FXS) ports on the router.


On-Net Calls
On-net calls occur between two telephones on the same data network, as shown in Figure 3-2. The calls can be routed through one or more Cisco voice-enabled routers, but the calls remain on the same data network. The edge telephones attach to the network through direct connections and FXS ports, or through a PBX, which typically connects to the network via a T1 connection. IP phones that connect to the network via switches place on-net calls through Cisco Unified CallManager. The connection across the data network can be a LAN connection, as in a campus environment, or a WAN connection, as in an enterprise environment.


Off-Net Calls
Figure 3-3 shows an example of an off-net call. To gain access to the public switched telephone network (PSTN), the user dials an access code, such as 9, from a telephone that is directly connected to a Cisco voice-enabled router or PBX. The connection to the PSTN is typically a single analog connection via a Foreign Exchange Office (FXO) port or a digital T1 or E1 connection.


PLAR Calls
PLAR calls automatically connect a telephone to a second telephone when the first telephone goes off hook, as depicted in Figure 3-4. When this connection occurs, the user does not get a dial tone because the voice-enabled port that the telephone is connected to is preconfigured with a specific number to dial. A PLAR connection can work between any types of signaling, including receive and transmit (ear and mouth [E&M]), FXO, FXS, or any combination of analog and digital interfaces. As an example, you might have encountered a PLAR connection at an airline ticket counter, where you pick up a handset and are immediately connected with an airline representative.


PBX-to-PBX Calls
PBX-to-PBX calls, as shown in Figure 3-5, originate at a PBX at one site and terminate at a PBX at another site while using the network as the transport between the two locations. Many business environments connect sites with private tie trunks. When migrating to a converged voice and data network, this same tie-trunk connection can be emulated across the IP network. Modern PBX connections to the network are typically digital T1 or E1 with channel associated signaling (CAS) or PRI signaling, although PBX connections can also be analog.


CallManager-to-CallManager Calls
As part of an overall migration strategy, a business might replace PBXs with a Cisco Unified CallManager infrastructure. This infrastructure includes IP telephones that

plug directly into the IP network. Cisco Unified CallManager performs the same call-routing functions formerly provided by the PBX. When an IP phone uses Cisco Unified CallManager to place a call, Cisco CallManager, based on its configuration, assesses whether the call is destined for another IP phone under its control or whether the call must be routed through a remote Cisco CallManager for call completion. Although the call stays on the IP network, it might be sent between zones. Every Cisco CallManager is part of a zone. A zone is a collection of devices that are under a common administration, usually a Cisco Unified CallManager or gatekeeper. Figure 3-6 provides an example of a CallManager-to-CallManager call.


On-Net to Off-Net Calls
When planning a resilient call-routing strategy, it might be necessary to reroute calls through a secondary path should the primary path fail. An on-net to off-net call, as illustrated in Figure 3-7, originates on an internal network and is routed to an external network, usually to the PSTN. On-net to off-net call-switching functionality might be necessary when a network link is down or if a network becomes overloaded and unable to handle all calls presented.

Common Channel Signaling Systems

Common channel signaling (CCS) differs from CAS in that all channels use a common channel and protocol for call setup. Using E1 as an example, a signaling protocol, such as ISDN Q.931, would be deployed in time slot 17 to exchange call-setup messages with its attached telephony equipment, as seen in Figure 2-36.


Examples of CCS signaling are as follows:

• Proprietary implementations Some PBX vendors choose to use CCS for T1 and E1 and implement a proprietary CCS protocol between their PBXs. In this implementation, Cisco devices are configured for Transparent Common Channel Signaling (T-CCS) because the Cisco devices do not understand proprietary signaling information.

• Integrated Services Digital Network (ISDN) ISDN uses Q.931 in a common channel to signal all other channels.

• Q Signaling (QSIG) Like ISDN, QSIG uses a common channel to signal all other channels.

• Digital Private Network Signaling System (DPNSS) DPNSS is an open standard developed by British Telecom for implementation by any vendor who chooses to use it. DPNSS also uses a common channel to signal all other channels.

• Signaling System 7 (SS7) SS7 is an out-of-band network implemented and maintained by various telephone companies and used for signaling and other supplemental services.

The following discussions elaborate on various CCS implementations. Note that proprietary implementations are not discussed because they vary widely among vendors.


ISDN

ISDN (Integrated Services Digital Network) is an access specification to a network. You may have studied ISDN as an access method for dial-up data systems. Because it is a digital system, ISDN makes connections rapidly.

ISDN can be implemented in two different ways: BRI (Basic Rate Interface) and PRI (Primary Rate Interface). BRI features two bearer (B) channels, while PRI supports 23 (for T1) or 30 (for E1) B channels. Each implementation also supports a data (D) channel, used to carry signaling information (CCS).

The following are benefits of using ISDN to transmit voice:

• Each B channel is 64 kbps, making it perfect for G.711 PCM.

• ISDN has a built-in call control protocol known as ITU-T Q.931.

• ISDN can convey standards-based voice features, such as call forwarding.

• ISDN supports standards-based enhanced dial-up capabilities, such as Group 4 fax and audio channels.

Figure 2-37 shows the architecture of an ISDN network. The B channel carries information, such as voice, data, and video, at 64 kbps. The D channel carries call signaling between customer premises equipment (CPE) and the network, usually as the Q.931 protocol but sometimes as the QSIG protocol.


BRI operates using the average local copper pair. It uses two B channels and one signaling channel, which is written as 2 B + D.

PRI implemented on T1 uses 23 B channels and one signaling channel, which is written as 23 B + D. PRI implemented on E1 uses 30 B channels and one signaling channel, which is represented as 30 B + D.

ISDN's Q.931 protocol, which operates at Layer 3 of the OSI (Open System Interconnection) model, uses a standard set of messages to communicate, as illustrated in Figure 2-38.


These standard messages cover the following areas:

• Call establishment Initially sets up a call. Messages travel between the user and the network. Call establishment events include alerting, call proceeding, connect, connect acknowledgment, progress, setup, and setup acknowledgment.

• Call information phase Data sent between the user and the network after the call is established. This allows the user to, for example, suspend and then resume a call. Events in the call information phase include hold, hold acknowledgment, hold reject, resume, resume acknowledgment, resume reject, retrieve, retrieve acknowledgment, retrieve reject, suspend, suspend acknowledgment, suspend reject, and user information.

• Call clearing Terminates a call. The following events occur in the call-clearing phase: disconnect, release, release complete, restart, and restart acknowledgment.

• Miscellaneous messages Negotiates network features (supplementary services). Miscellaneous services include congestion control, facility, information, notify, register, status, and status inquiry.

QSIG

The QSIG (Q Signaling) protocol is based on the ISDN Q.931 standard and provides signaling for private integrated services network exchange (PINX) devices. Figure 2-39 shows how different QSIG operations map to the OSI model.


DPNSS

British Telecom and selected PBX manufacturers originally developed the Digital Private Network Signaling System (DPNSS) in the early 1980s. It was developed and put into use before the ISDN standards were completed because customers wanted to make use of digital facilities as soon as possible.

DPNSS operates over standard ISDN physical interfaces and is described in four documents:

• BTNR 188: Digital Private Networking Signalling System No 1, Issue 6, January 1995.

• BTNR 188-T: Digital Private Networking Signalling System No 1: Testing Schedule.

• BTNR 189: Interworking between DPNSS1 and other Signalling Systems, Issue 3, March 1988.

• BTNR 189-I: Interworking between DPNSS1 and ISDN Signalling Systems, Issue 1, December 1992.

SIGTRAN

SIGTRAN, as illustrated in Figure 2-40, is a signaling protocol defined in RFC 2719 and RFC 2960. It describes the way the IP protocol carries SS7 messages in a VoIP network. SIGTRAN relies on the Stream Control Transport Protocol at Layer 4 of the TCP/IP protocol stack.


Using SIGTRAN, a service provider may interconnect a private VoIP network to the public switched telephone network (PSTN) and ensure that SS7 signals are conveyed end to end.

Signaling Systems

Configuring Cisco Systems voice equipment to interface with other equipment requires an understanding of the signaling that conveys supervision between the systems. Proper troubleshooting also requires an understanding of these signaling systems.

This section describes the various signaling systems used between telephony systems, such as common channel signaling and channel associated signaling. It also explores signaling between PBXs, signaling between PBXs and COs, and specialized signaling, such as ISDN.


Channel Associated Signaling

Channel associated signaling (CAS) is a signaling method commonly used between PBXs. Although this can manifest itself in many forms, some methods are more common than others. Signaling systems can also be implemented between a PBX and a Cisco voice device.


T1 Channel Associated Signaling

PBXs and Cisco devices use T1 and E1 interfaces to convey voice. Originally, this was the main purpose of T1, which carries signaling information using two methodologies: CAS and common channel signaling (CCS). Figure 2-31 illustrates the format of the T1 digital signal.


The characteristics of the T1 digital signal format are as follows:

• A T1 frame is 193 bits long, 8 bits from each of the 24 time slots (digital service zeros [DS0s]), plus 1 bit for framing. A T1 repeats every 125 microseconds, resulting in 8000 samples per second (8 bits * 24 time slots + 1 framing bit * 8000 samples per second = 1.544 Mbps).
  

• T1 has two major framing and format standards:

- Super Frame (SF), or D4, specifies 12 frames in sequence. The D4 framing pattern used in the F position in Figure 2-31 is 100011011100 (a 1 goes with the first frame, a 0 goes with the second frame, a 0 goes with the third frame, and so on, all the way through 12 frames). This unique framing pattern allows the receiving T1 equipment to synchronize within four frames, since any four consecutive frame bits are unique within the 12-bit pattern. Because there are 8000 T1 frames transmitted per second, 8000 F bits are produced and used for framing.

- Extended Superframe (ESF) format was developed as an upgrade to SF and is now dominant in public and private networks. Both types of formatting retain the basic frame structure of one framing bit followed by 192 data bits. However, ESF repurposes the use of the F bit. In ESF, of the total 8000 F bits used in T1, 2000 are used for framing, 2000 are used for cyclic redundancy check (CRC) (for error checking only), and 4000 are used as an intelligent supervisory channel to control functions end to end (such as loopback and error reporting).

Because each DS0 channel carries 64 kbps, and G.711 is 64 kbps, there is no room to carry signaling. Implemented for voice, the T1 uses every sixth frame to convey signaling information. In every sixth frame, the least significant bit (LSB) for each of the voice channels is used to convey the signaling, as shown in Figure 2-32. Although this implementation detracts from the overall voice quality (because only 7 bits represent a sample for that frame), the impact is not significant. This method is called robbed-bit signaling (RBS). When SF employs this method, the signaling bits are conveyed in both the 6th (called the "A" bit) and 12th (called the "B" bit) frames. For control signaling, A and B bits provide both near- and far-end off-hook indication.


The A and B bits can represent different signaling states or control features (on hook or off hook, idle, busy, ringing, and addressing). The robbed bit is the least significant bit from an 8-bit word.

ESF also uses RBS in frames 6, 12, 18, and 24 to yield four signaling bits, providing additional control and signaling information. These four bits are known as the A, B, C, and D bits.

Because the signaling occurs within each DS0, it is referred to as in band. Also, because the use of these bits is exclusively reserved for signaling each respective voice channel, it is referred to as CAS.

The robbed bits, depicted in Figure 2-33, are used to convey E&M status or FXS/FXO status and provide call supervision for both on hook and off hook.


T1 CAS has the following characteristics:

• SF has a 12-frame structure and provides AB bits for signaling.

• ESF has a 24-frame structure and provides ABCD bits for signaling.

• DTMF, or tone, can be carried in band in the audio path. However, other supervisory signals must still be carried via CAS.

E1 Channel Associated Signaling

In E1 framing and signaling, 30 of the 32 available channels, or time slots, are used for voice and data. Framing information uses time slot 1, while time slot 17 (E0 16) is used for signaling by all the other time slots. This signaling format, illustrated in Figure 2-34, is also known as CAS because the use of the bits in the 17th time slot is exclusively reserved for the purpose of signaling each respective channel. However, this implementation of CAS is considered out of band, because the signaling bits are not carried within the context of each respective voice channel, as is the case with T1. E1 CAS is directly compatible with T1 CAS, because both methods use AB or ABCD bit signaling. Although the signaling for E1 CAS is carried in a single common time slot, it is still referred to as CAS because each individual signaling time slot represents a specific pair of voice channels.


In the E1 frame format, 32 time slots make up a frame. A multiframe consists of 16 E1 frames, as depicted in Figure 2-35.

Voice Compression Standards

To conserve valuable WAN bandwidth, you can compress the quantized voice waveforms. Two categories of waveform encoding include:

• Waveform algorithms (coders) Waveform algorithms have the following functions and characteristics:

- Sample analog signals at 8000 times per second

- Use predictive differential methods to reduce bandwidth

- Highly impact voice quality because of reduced bandwidth

- Do not take advantage of speech characteristics

- Examples include: G.711 and G.726

• Source algorithms (coders) Source algorithms have the following functions and characteristics:

- Source algorithm coders are called vocoders, or voice coders. A vocoder is a device that converts analog speech into digital speech, using a specific compression scheme that is optimized for coding human speech.

- Vocoders take advantage of speech characteristics.

- Bandwidth reduction occurs by sending linear-filter settings.

- Codebooks store specific predictive waveshapes of human speech. They match the speech, encode the phrases, decode the waveshapes at the receiver by looking up the coded phrase, and match it to the stored waveshape in the receiver codebook.

- Examples include: G.728 and G.729


The following three common voice compression techniques are standardized by the ITU-T:

• PCM Amplitude of voice signal is sampled and quantized at 8000 times per second. Each sample is then represented by one octet (8 bits) and transmitted. For sampling, you must use either a-law or µ-law to reduce the signal-to-noise ratio.

• ADPCM The difference between the current sample and its predicted value (based on past samples). ADPCM is represented by 2, 3, 4, or 5 bits. This method reduces the bandwidth requirement at the expense of signal quality.

• CELP Excitation value and a set of linear-predictive filters (settings) are transmitted. The filter setting transmissions are less frequent than excitation values and are sent on an as-needed basis.

Table 2-4 describes the CODECs and compression standards.

A common type of waveform encoding is pulse code modulation (PCM). Standard PCM is known as ITU standard G.711, which requires 64,000 bits per second of bandwidth to transport the voice payload (that is, not including any overhead), as shown in Figure 2-30.


Figure 2-30 shows that PCM requires 1 polarity bit, 3 segment bits, and 4 step bits, which equals 8 bits per sample. The Nyquist Theorem requires 8000 samples per second; therefore, you can figure the required bandwidth as follows:

8 bits * 8000 samples per second = 64,000 bits per second

Adaptive differential pulse code modulation (ADPCM) coders, like other waveform coders, encode analog voice signals into digital signals to adaptively predict future encodings by looking at the immediate past. The adaptive feature of ADPCM reduces the number of bits per second that the PCM method requires to encode voice signals. ADPCM does this by taking 8000 samples per second of the analog voice and turning them into linear PCM samples. ADPCM then calculates the predicted value of the next sample, based on the immediate past sample, and encodes the difference. The ADPCM process generates 4-bit words, thereby generating 16 specific bit patterns.

The ADPCM algorithm from the Consultative Committee for International Telegraph and Telephone (CCITT) transmits all 16 possible bit patterns. The ADPCM algorithm from the American National Standards Institute (ANSI) uses 15 of the 16 possible bit patterns. The ANSI ADPCM algorithm does not generate a 0000 pattern.

The ITU standards for compression are as follows:

• G.711 rate: 64 kbps = (2 * 4 kHz) * 8 bits/sample

• G.726 rate: 32 kbps = (2 * 4 kHz) * 4 bits/sample

• G.726 rate: 24 kbps = (2 * 4 kHz) * 3 bits/sample

• G.726 rate: 16 kbps = (2 * 4 kHz) * 2 bits/sample

Code excited linear prediction (CELP) compression transforms analog voice as follows:

1. The input to the coder is converted from an 8-bit to a 16-bit linear PCM sample.

2. A codebook uses feedback to continuously learn and predict the voice waveform.

3. The coder is excited (that is, begins its lookup process) by a white noise generator.

4. The mathematical result is sent to the far-end decoder for synthesis and generation of the voice waveform.

Two forms of CELP include Low-Delay CELP (LDCELP) and Conjugate Structure Algebraic CELP (CS-ACELP). LDCELP is similar to CS-ACELP, except for the following:

• LDCELP uses a smaller codebook and operates at 16 kbps to minimize delay, or look-ahead, from 2 to 5 ms, while CS-ACELP minimizes bandwidth requirements (8 kbps) at the expense of increasing delay (10 ms).

• The 10-bit code word is produced from every five speech samples from the 8 kHz input with no look-ahead.

• Four of these 10-bit code words are called a subframe. They take approximately 2.5 ms to encode. CS-ACELP uses eight 10-bit code words.

Two of these subframes are combined into a 5-ms block for transmission. CS-ACELP is a variation of CELP that performs these functions:

• Codes 80-byte frames, which take approximately 10 ms to buffer and process.

• Adds a look-ahead of 5 ms. A look-ahead is a coding mechanism that continuously analyzes, learns, and predicts the next waveshape.

• Adds noise reduction and pitch-synthesis filtering to processing requirements.

Cisco VoIP environments typically leverage the benefits of G.729 when transmitting voice traffic over the IP WAN. These benefits include the ability to minimize bandwidth demands, while maintaining an acceptable level of voice quality. Several variants of G.729 exist.

Analog-to-Digital and Digital-to-Analog Voice Encoding

This section covers the fundamentals of digitally encoding voice, specifically, the basics of voice digitization and the various compression schemes that are used to transport voice while using less bandwidth.

Digitizing speech was a project first undertaken by the Bell System in the 1950s. The original purpose of digitizing speech was to deploy more voice circuits with a smaller number of wires. This evolved into the T1 and E1 transmission methods of today. Examples of analog and digital waveforms are presented in Figure 2-24.


Table 2-3 details the steps to convert an analog signal to a digital signal.


The three mandatory components in the analog-to-digital conversion process are further described as follows:

• Sampling Sample the analog signal at periodic intervals. The output of sampling is a pulse amplitude modulation (PAM) signal.

• Quantization Match the PAM signal to a segmented scale. This scale measures the amplitude (height) of the PAM signal and assigns an integer number to define that amplitude.

• Encoding Convert the integer base-10 number to a binary number. The output of encoding is a binary expression in which each bit is either a 1 (pulse) or a 0 (no pulse).

This three-step process is repeated 8000 times per second for telephone voice-channel service. Use the fourth optional step, compression, to save bandwidth. This optional step allows a single channel to carry more voice calls.

After the receiving terminal at the far end receives the digital PCM signal, it must convert the PCM signal back into an analog signal. The process of converting digital signals back into analog signals includes the following two processes:

• Decoding The received 8-bit word is decoded to recover the number that defines the amplitude of that sample. This information is used to rebuild a PAM signal of the original amplitude. This process is simply the reverse of the analog-to-digital conversion.

• Filtering The PAM signal is passed through a filter to reconstruct the original analog wave form from its digitally coded counterpart.

With this basic understanding of analog to digital conversion, this chapter considers the sampling, quantization, and encoding processes more thoroughly, beginning with sampling.


Sampling and the Nyquist Theorem

One of the major issues with sampling is determining how often to take those samples (that is, "snapshots") of the analog wave. You do not want to take too few samples per second because when the equipment at the other end of the phone call attempts to reassemble and make sense of those samples, a different sound (that is, a lower frequency sound) signal might also match those samples, and the incorrect sound would be heard by the listener. This phenomenon is called aliasing, as shown in Figure 2-25.


With the obvious detrimental effect of undersampling, you might be tempted to take many more samples per second. While that approach, sometimes called oversampling, does indeed eliminate the issue of aliasing, it also suffers from a major drawback. If you take far more samples per second than actually needed to accurately recreate the original signal, you consume more bandwidth than is absolutely necessary. Because bandwidth is a scarce commodity (especially on a wide-area network), you do not want to perform the oversampling shown in Figure 2-26.


Digital signal technology is based on the premise stated in the Nyquist Theorem: When a signal is instantaneously sampled at the transmitter in regular intervals and has a rate of at least twice the highest channel frequency, then the samples will contain sufficient information to allow an accurate reconstruction of the signal at the receiver. Figure 2-27 illustrates sampling, as prescribed by the Nyquist Theorem.


While the human ear can sense sounds from 20 to 20,000 Hz, and speech encompasses sounds from about 200 to 9000 Hz, the telephone channel was designed to operate at about 300 to 3400 Hz. This economical range carries enough fidelity to allow callers to identify the party at the far end and sense their mood. Nyquist decided to extend the digitization to 4000 Hz, to capture higher-frequency sounds that the telephone channel may deliver. Therefore, the highest frequency for voice is 4000 Hz, or 8000 samples per second; that is, one sample every 125 microseconds.


Quantization

Quantization involves dividing the range of amplitude values that are present in an analog signal sample into a set of discrete steps that are closest in value to the original analog signal, as illustrated in Figure 2-28. Each step is assigned a unique digital code word.


In Figure 2-28, the x-axis is time and the y-axis is the voltage value (PAM). The voltage range is divided into 16 segments (0 to 7 positive, and 0 to 7 negative). Starting with segment 0, each segment has fewer steps than the previous segment, which reduces the signal-to-noise ratio (SNR) and makes the segment uniform. This segmentation also corresponds closely to the logarithmic behavior of the human ear. If there is an SNR problem, it is resolved by using a logarithmic scale to convert PAM to PCM.

Linear sampling of analog signals causes small-amplitude signals to have a lower SNR, and therefore poorer quality, than larger amplitude signals. The Bell System developed the µ-law method of quantization, which is widely used in North America. The International Telecommunication Union (ITU) modified the original m-law method and created a-law, which is used in countries outside of North America.

By allowing smaller step functions at lower amplitudes, rather than higher amplitudes, µ-law and a-law provide a method of reducing this problem. Both µ-law and a-law "compand" the signal; that is, they both compress the signal for transmission and then expand the signal back to its original form at the other end.

Informational Signaling

DTMF tones are used not just for address signaling but also for informational signaling. Specifically, call-progress indicators in the form of tone combinations are used to notify subscribers of call status. Each combination of tones represents a different event in the call process, as follows:

• Dial tone Indicates that the telephone company is ready to receive digits from the user telephone. Cisco routers provide dial tone as a method of showing that the hardware is installed. In a PBX or key telephone system, the dial tone indicates the system is ready to receive digits.

• Busy Indicates that a call cannot be completed because the telephone at the remote end is already in use.

• Ringback (CO or PBX) Indicates that the telephone switch is attempting to complete a call on behalf of a subscriber.

• Congestion Indicates that congestion in the long-distance telephone network is preventing a telephone call from being processed. The congestion tone is sometimes known as the all-circuits-busy tone.

• Reorder Indicates that all of the local telephone circuits are busy, thus preventing a telephone call from being processed. The reorder tone is known to the user as fast-busy and is familiar to anyone who operates a telephone from a PBX.

• Receiver off hook Indicates that the receiver has been off hook for an extended period without placing a call.

• No such number Indicates that a subscriber placed a call to a nonexistent number.

Trunk Connections

Before a telephone call terminates at its final destination, the call is routed through multiple switches. When a switch receives a call, it determines whether the destination telephone number is within a local switch or if the call needs to go through another switch to a remote destination. Trunks interconnect the telephone company and PBX switches, as shown in Figure 2-9.

The primary function of the trunk is to provide the path between switches. The switch must route the call to the correct trunk or telephone line. Although many different subscribers share a trunk, only one subscriber uses it at any given time. As telephone calls end, they release trunks and make them available to the switch for subsequent calls. There can be several trunks between two switches.

The following are examples of the more common trunk types:

• Private trunk lines (tie-lines) Companies with multiple PBXs often connect them with tie trunk lines. Generally, tie trunk lines serve as dedicated circuits that connect PBXs. On a monthly basis, subscribers lease trunks from the telephone company to avoid the expense of using telephone lines on a per-extension basis. These types of connections, known as tie-lines, typically use special interfaces called recEive and transMit, or E&M interfaces.

• CO trunks A CO trunk serves as a direct connection between a PBX and the local CO that routes calls; for example, the connection from a private office network to the public switched telephone network (PSTN). When users dial 9, they are connecting through their PBX to the CO trunk to access the PSTN. CO trunks typically use Foreign Exchange Office interfaces. Certain specialized CO trunks are frequently used on the telephony network. A direct inward dial trunk, for example, allows outside callers to reach specific internal destinations without having to be connected via an operator.

• Interoffice trunks An interoffice trunk is a circuit that connects two local telephone company COs.

• Foreign exchange (FX) trunks FX trunks are interfaces that are connected to switches supporting connections to either office equipment or station equipment. Office equipment includes other switches (to extend the connection) and Cisco devices. Station equipment includes telephones, fax machines, and modems.

Trunk Signaling

Lines and trunks must adhere to signaling standards just as telephony networks and telephone companies do. Trunk signaling serves to initiate the connection between the switch and the network. There are five different types of trunk signaling, and each applies to different kinds of interfaces, such as FXS, FXO, and E&M:

• Loop-start signaling

• Ground-start signaling

• E&M wink-start signaling

• E&M immediate-start signaling

• E&M delay-start signaling

The following sections explain these signaling types.


Loop-Start Signaling

Loop-start signaling allows a user or the telephone company to seize a line or trunk when a subscriber is initiating a call. It is primarily used on local loops connecting to residences rather than on trunks interconnecting telephone switches.

A telephone connection exists in one of the following states, as illustrated in Figure 2-10:

• Idle (on hook)
• Telephone seizure (off hook)
• CO seizure (ringing)

A summary of the loop-start signaling process is as follows:

1. When the line is in the idle state, or on hook, the telephone or PBX opens the two-wire loop. The CO or FXS has battery on ring and ground on tip.

2. If a user lifts the handset off the cradle to place a call, the switch hook goes off hook and closes the loop (line seizure). The current can now flow through the telephone circuit. The CO or FXS module detects the current and returns a dial tone.

3. When the CO or FXS module detects an incoming call, it applies AC ring voltage superimposed over the 48 VDC battery, causing the ring generator to notify the recipient of a telephone call. When the telephone or PBX answers the call, thus closing the loop, the CO or FXS module removes the ring voltage.

Loop-start signaling is a poor solution for high-volume trunks because it leads to glare, which is the simultaneous seizure of the trunk from both ends. Glare occurs, for example, when you pick up your home telephone and find that someone is already at the other end.

Glare is not a significant problem at home. It is, however, a major problem when it occurs between switches at high-volume switching centers, such as long-distance carriers or large PBX systems.


Ground-Start Signaling

Ground-start signaling, illustrated in Figure 2-11, is a modification of loop-start signaling that corrects for the probability of glare. It solves the problem by providing current detection at both ends.

Although loop-start signaling works when you use your telephone at home, ground-start signaling is preferable when there are high-volume trunks involved at telephone switching centers. Because ground-start signaling uses a request or confirm switch at both ends of the interface, it is preferable over other signaling methods on high-usage trunks, such as FXOs. FXOs require implementation of answer supervision (reversal or absence of current) on the interface for the confirmation of on hook or off hook.


E&M Signaling

E&M signaling supports tie-line type facilities or signals between voice switches. Instead of superimposing both voice and signaling on the same wire, E&M uses separate paths, or leads, for each.

To call a remote office, your PBX must route a request for use of the trunk over its signal leads between the two sites. Your PBX makes the request by activating its M-lead. The other PBX detects the request when it detects current flowing on its E-lead. It then attaches a dial register to the trunk and your PBX, which sends the dialed digits. The remote PBX activates its M-lead to notify the local PBX that the call has been answered.

There are five types of E&M signaling: Type I, Type II, Type III, Type IV, and Type V. The E&M leads operate differently with each wiring scheme, as shown in Table 2-1 and Table 2-2. Keep in mind that any of the E&M supervisory signaling types (that is, wink-start, immediate-start, and delay-start) can operate over any of the following wiring schemes.

Analog and Digital Voice Connections

Analog Voice Fundamentals

Interfacing Cisco Systems equipment with traditional analog telephony devices requires an understanding of the various interfaces used in the industry. This section introduces analog interfaces that you can select from, including Foreign Exchange Station (FXS), Foreign Exchange Office (FXO), and ear and mouth (E&M).


Local-Loop Connections

A subscriber home telephone connects to the telephone company central office (CO) via an electrical communication path called a local loop, as illustrated in Figure 2-1. The loop consists of a pair of twisted wires. One is called tip; the other is called ring, as shown in Figure 2-2.


The names tip and ring come from the plug used by the operators of yesteryear to interconnect calls. As you can see in Figure 2-2, the plug used by these operators resembles the plug you might use to connect your headphones to your home stereo equipment. There are three conductors on this plug. The conductor (that is, wire) connected to the tip of the plug is called the tip wire, and the conductor connected to the ring in the middle of the plug is called the ring wire.

In most arrangements, the ring wire ties to the negative side of a power source, called the battery, while the tip wire connects to the ground. When you take your telephone off hook, current flows around the loop, allowing dial tone to reach your handset. Your local loop, along with all others in your neighborhood, connects to the CO in a cable bundle, either buried underground or strung on poles.

Note
Your home telephone service is provided to you from your service provider by way of two wires. Your home telephone controls whether the service on these wires is activated via the switch hook inside the telephone.


Local-Loop Signaling

A subscriber and telephone company notify each other of the call status through audible tones and an exchange of electrical current. This exchange of information is called local-loop signaling. Local-loop signaling consists of supervisory signaling, address signaling, and informational signaling, each of which has its own characteristics and purpose. The three types of local-loop signaling appear on the local loop and serve to prompt the subscriber and the switch into a certain action.


Supervisory Signaling

Resting the handset on the telephone cradle opens the switch hook and prevents the circuit current from flowing through the telephone, as seen in Figure 2-3. Regardless of the signaling type, a circuit goes on hook when the handset is placed on the telephone cradle and the switch hook is toggled to an open state. When the telephone is in this position, only the ringer is active.


To place a call, a subscriber must lift the handset from the telephone cradle. Removing the handset from the cradle places the circuit off hook, as shown in Figure 2-4. The switch hook is then toggled to a closed state, causing circuit current to flow through the electrical loop. The current notifies the telephone company that someone is requesting to place a telephone call. When the telephone network senses the off-hook connection by the flow of current, it provides a signal in the form of the dial tone to indicate that it is ready.


When a subscriber makes a call, the telephone switch sends voltage to the ringer to notify the other subscriber of an inbound call, as illustrated in Figure 2-5. The telephone company also sends a ringback tone to the caller, alerting the caller that it is sending ringing voltage to the recipient telephone.


The pattern of the ring signal, or ring cadence, varies around the world. As depicted in Figure 2-6, the ring cadence (that is, ringing pattern) in the United States is 2 seconds of ringing followed by 4 seconds of silence. The United Kingdom uses a double ring of 0.4 seconds separated by 0.2 seconds of silence, followed by 2 seconds of silence.

Address Signaling

Although somewhat outdated, rotary-dial telephones are still in use and easily recognized by their large numeric dial-wheel. When placing a call, the subscriber spins the large numeric dial-wheel to send digits. These digits must be produced at a specific rate and within a certain level of tolerance. Each pulse consists of a "break" and a "make," as detailed in Figure 2-7. The break segment is the time that the circuit is open. The make segment is the time during which the circuit is closed. In the United States, the break-and-make cycle must correspond to a ratio of 60 percent break to 40 percent make.


A governor inside the dial controls the rate at which the digits are pulsed. The dial pulse signaling process occurs as follows:

1. When a subscriber calls someone by dialing a digit on the rotary dial, a spring winds.

2. the dial is released, the spring rotates the dial back to its original position.


3. While the spring rotates the dial back to its original position, a cam-driven switch opens and closes the connection to the telephone company. The number of consecutive opens and closes (that is, breaks and makes) represents the dialed digit.


A more modern approach to address signaling is touch-tone dialing. Users who have a touch-tone pad or a push-button telephone must push the keypad buttons to place a call, rather than rotating a dial as they did with pulse dialing. Each button on the keypad is associated with a set of high and low frequencies. Each row of keys on the keypad is identified by a low-frequency tone; each column of keys on the keypad is identified by a high-frequency tone. The combination of both tones notifies the telephone company of the number being called, hence the term dual-tone multifrequency (DTMF). Figure 2-8 illustrates the combination of tones generated for each button on the keypad.

IP Telephony Applications

Types of Deployment

When deploying VoIP technologies, design decisions should take into account the environment in which VoIP is being installed. This section considers three typical environments: the campus LAN, enterprise, and service provider environments.


Campus LAN Environment

Campus LAN environments, an example of which is illustrated in Figure 1-24, have grown tremendously in the past several years due to the demand for networked resources, instant business communication, and VoIP applications.


Components for integrated voice and data campus networks, as discussed previously in the "Packet Telephony Components" section, include the following:

• IP Phone

• Gateway

• MCU

• Application server

When you are designing the campus infrastructure for voice, you must consider the following key issues:

• Robust, fault-tolerant, highly available network design

• Ability to power IP phones

• Redundant power supply for network components

• Ease of IP addressing

• QoS enhancements

Cisco Systems' internal telephone network in San Jose can be considered a campus LAN environment. All desktop phones connect to Ethernet switches and are controlled by Unified CallManager applications. Unified CallManager also controls the gateways and other application servers, such as the Unity server.


Enterprise Environment

Enterprise networks grow and evolve as company services and locations change and expand. Heavy reliance on information processing and universal access to corporate information has driven network designs to provide reliable access, redundancy, reachability, and manageability. These same principles apply to designing corporation-wide voice access in the enterprise environment.

Enterprise networks can be either centralized or distributed call processing environments. In the centralized call processing environment, all of the components of the voice system are controlled by a single centralized call agent, such as Unified CallManager, regardless of their physical location. In a distributed call processing environment, the components of the voice network at each location can act independently.

Figure 1-25 depicts an enterprise centralized call processing environment. Centralized voice networks provide enterprise-wide voice access for calls and voice services controlled from a central site. In this environment, the central site provisions all voice services, such as Cisco Unified CallManager, voice mail, and unified messaging. IP phones at remote sites connect to Cisco Unified CallManager through the IP WAN for call processing.


Components for centralized voice enterprise networks include the following:

• IP phone

• Cisco Unified CallManager cluster (central site only)

• Gateway (all sites)

• MCU (central site only)

• Application server (central site only)

• SRST based on Cisco IOS software (remote sites only)

• IP WAN

Figure 1-26 shows an enterprise distributed call processing environment. Distributed voice networks place voice components at each site and utilize the WAN for intersite calls only.


Components for distributed voice enterprise networks include the following:

• IP phone

• Cisco Unified CallManager cluster

• Gateway

• MCU

• Application server (all sites)

• IP WAN

Modern enterprise network applications include:

• E-business

• E-learning

• Customer care

• Unified messaging

• Videoconferencing

• Voice calls placed from web pages

Service Provider Environment

Service provider environments, an example of which is illustrated in Figure 1-27, add another level of complexity to the voice environment. To be competitive, service providers must provide their business customers with more efficient, less expensive alternatives to the PSTN for voice and data services.


Requirements in the service provider arena include:

• Carrier class performance Voice gateways must provide service that minimizes latency and controls jitter. This level of performance allows customers to maintain voice quality as they migrate from circuit-switched voice to IP-based services.

• Scalability Design must accommodate rapid growth to enable service providers to grow with their customer base. An important aspect of scalability is the automation, configuration, and administration of IP networks and gateways for seamless expansion.

• Comprehensive call records supporting flexible service pricing This is the ability to extract IP session and transaction information from multiple network devices and from all layers of the network, in real time, to produce detailed billing records.

• Signaling System 7 (SS7) interconnect capabilities Tariffs favor interconnection using SS7 signaling because Inter-Machine Trunks (IMTs) are less expensive than ISDN-based facilities. This financial benefit equates to lower monthly expenses, reduced cost of goods that are sold, and higher margins for service providers.

IP Telephony Applications

As customers migrate their voice networks, they face a myriad of choices regarding interface types, components, and topologies. A good network design incorporates solutions for current requirements and allows room for future growth. It is important to understand how voice interfaces with a network and how the components fit together to provide service in any environment.


Analog Interfaces

A Foreign Exchange Station (FXS) interface, as depicted in Figure 1-16, provides a direct connection to an analog telephone, a fax machine, or a similar device. From a telephone perspective, the FXS interface functions like a telephone switch (for example, a PBX); therefore, it must supply line power, ring voltage, and dial tone.


The FXS interface contains the coder-decoder (CODEC), which converts the spoken analog voice wave into a digital format for processing by the voice-enabled device.

The Foreign Exchange Office (FXO) interface, shown in Figure 1-17, allows an analog connection to be directed at the CO of a PSTN or to a station interface on a PBX. The switch recognizes the FXO interface as a telephone because the interface plugs directly into the line side of the switch. The FXO interface provides either pulse or DTMF digits for outbound dialing.


In PSTN terminology, an FXO-to-FXS connection is also referred to as a foreign exchange (FX) trunk. An FX trunk is a CO trunk that has access to a distant CO. Because this connection is FXS at one end and FXO at the other end, it acts as a long-distance extension of a local telephone line. In this instance, a local user can pick up the telephone and get a dial tone from a foreign city. Users in the foreign city can dial a local number and have the call connect to the user in the local city.

The E&M interface, shown in Figure 1-18, provides signaling for analog trunking. Analog trunk circuits connect automated systems (PBXs) and networks (COs). E&M signaling is also referred to as "ear and mouth," but its origin comes from the term "Earth and Magneto." Earth represents the electrical ground, and magneto represents the electromagnet used to generate tone.


E&M signaling defines a trunk-circuit side and a signaling-unit side for each connection, similar to the DCE and DTE reference types. The PBX is usually the trunk-circuit side, and the telco, CO, channel bank, or Cisco voice-enabled platform is the signaling-unit side.


Digital Interfaces

In a corporate environment with a large volume of voice traffic, connections to the PSTN and to PBXs are primarily digital. Examples of digital interfaces include T1, E1, and BRI interfaces.

T1 Interface

A T1 interface, as illustrated in Figure 1-19, is a form of digital connection that can simultaneously carry up to 24 conversations using two-wire pairs. When a T1 link operates in full-duplex mode, one wire pair sends and the other wire pair receives. The 24 channels are grouped together to form a frame. The frames are then grouped together into Super Frames (groups of 12 frames) or into Extended Superframes (groups of 24 frames).


The T1 interface carries either CAS or CCS. When a T1 interface uses CAS, the signaling robs a sampling bit for each channel to convey in band. When a T1 interface uses CCS, Q.931 signaling is used on a single channel, typically the last channel.

To configure CAS, you must specify the type of signaling that the robbed bits carry (for example, E&M Wink Start). This signaling must match the PSTN requirements or the PBX configuration. This is considered in-band signaling because the signal shares the same channel as the voice.

To configure CCS, you must configure the interface for PRI signaling. This level of configuration makes it possible to use channels 1 to 23 (called B channels) for voice traffic. Channel 24 (called the D channel) carries the Q.931 call control signaling for call setup, maintenance, and teardown. This type of signaling is considered out-of-band signaling because the Q.931 messages are sent in the D channel only.


E1 Interface

An E1 interface, shown in Figure 1-20, has 32 channels and simultaneously carries up to 30 conversations. The other two channels are used for framing and signaling. The 32 channels are grouped to form a frame. The frames are then grouped together into multiframes (groups of 16 frames). In Europe and Mexico, the E1 interface is most often used, while in the United States the T1 interface is most commonly used.


Although you can configure the E1 interface for either CAS or CCS, the most common usage is CCS.

When an E1 interface uses CAS, signaling travels out of band in the signaling channel but follows a strict association between the signal carried in the signaling channel and the channel to which the signaling is being applied. The signaling channel is channel 16.

In the first frame, channel 16 carries 4 bits of signaling for channel 1 and 4 bits of signaling for channel 17. In the second frame, channel 16 carries 4 bits of signaling for channel 2 and 4 bits for channel 18, and so on. This process makes it out-of-band CAS.

When an E1 interface uses CCS, Q.931 signaling is used on a single channel, typically channel 17. When configuring for CCS, configure the interface for PRI signaling. When E1 is configured for CCS, channel 16 carries Q.931 signaling messages only.


BRI Interface

Figure 1-21 depicts an Integrated Services Digital Network (ISDN) Basic Rate Interface (BRI). You can use a BRI to connect the PBX voice into the network. Used primarily in Europe for PBX connectivity, BRI provides a 16-kbps D channel for signaling and two 64-kbps B channels for voice. BRI uses Q.931 signaling in the D channel for call signaling.



IP Phones

Figure 1-22 depicts physical connection options for IP phones. The IP phone connects to the network through a Category 5 or better cable that has RJ-45 connectors. The power-enabled switch port or an external power supply provides power to an IP phone. The IP phone functions like other IP-capable devices sending IP packets to the IP network. Because these packets are carrying voice, you must consider both logical and physical configuration issues.

At the physical connection level, there are three options for connecting the IP phone:

• Single cable A single cable connects the telephone and the PC to the switch. Most enterprises install IP phones on their networks using a single cable for both the telephone and a PC. Reasons for using a single cable include ease of installation and cost savings on cabling infrastructure and wiring-closet switch ports.

• Multiple cables Separate cables connect the telephone and the PC to the switch. Users often connect the IP phone and PC using separate cables. This connection creates a physical separation between the voice and data networks.

• Multiple switches Separate cables connect the telephone and the PC to separate switches. With this option, IP phones are connected to separate switches in the wiring closet. By using this approach, you can avoid the cost of upgrading the current data switches and keep the voice and data networks completely separate.

Multiple switches are used to do the following:

• Provide inline power to IP phones without having to upgrade the data infrastructure

• Reduce the amount of Cisco IOS Catalyst software upgrades needed in the network

• Limit the spanning-tree configuration in the wiring-closet switches

The physical configuration for connecting an IP phone must address the following issues:

• Speed and duplex settings

• Inline power settings

The logical configuration for connecting an IP phone must address the following issues:

• IP addressing
  

• VLAN assignment

• Spanning tree

• Classification and queuing

Many Cisco IP phones, such as the 7970G shown in Figure 1-23, contain a three-port 10/100 switch. One port is an internal port that connects the voice electronics in the telephone. A second port connects a daisy-chained PC, and the third port uplinks to the Ethernet switch in the wiring closet.


If a computer is connected to an IP phone, data packets traveling to and from the computer, and to and from the phone, share the same physical link to the access layer switch and the same port on the access layer switch. This shared physical link has the following implications for the VLAN network configuration:

• Current VLANs may be configured on an IP subnet basis. However, additional IP addresses may not be available for assigning the telephone to the same subnet as the other devices that are connected to the same port.
  

• Data traffic that is supporting phones on the VLAN may reduce the quality of VoIP traffic.

You can resolve these issues by isolating the voice traffic on a separate VLAN for each of the ports connected to a telephone. The switch port configured for connecting a telephone would have separate VLANs configured to carry the following types of traffic:

• Voice traffic to and from the IP phone (auxiliary VLAN)

• Data traffic to and from the PC connected to the switch through the IP phone access port (native VLAN)

Fundamentals of Packet Telephony Networks

The increased efficiency of packet networks (for example, VoIP networks) and the ability to statistically multiplex voice traffic with data packets allows companies to maximize their return on investment (ROI) in data network infrastructures. Multiplexing voice traffic with data traffic reduces the number of costly circuits dedicated to servicing voice applications.

As demand for voice services expands, it is important to understand the different requirements of voice and data traffic. Previously, voice and data networks were separate and could not impact each other. Today, it is necessary to determine the protocols available to control voice calls and ensure that data flows are not negatively impacted.

This section delves into the benefits of packet telephony networks and provides an overview of basic packet telephony operations. Additionally, the fundamental components of packet networks are introduced. Finally, as a design consideration, this section considers the fragile nature of voice packets.


Packet Telephony Components

The basic components of a packet voice network, as shown in Figure 1-12, include the following:

• IP phones Provide IP voice to the desktop.

• Gatekeeper Provides Call Admission Control (CAC), bandwidth control and management, address translation, and call routing.

• Gateway Provides translation between VoIP and non-VoIP networks, such as the PSTN. It also provides physical access for local analog and digital voice devices, such as telephones, fax machines, key sets, and PBXs.

• Multipoint control unit (MCU) Provides real-time connectivity for participants in multiple locations to attend the same videoconference or meeting.

• Call agent Provides call control for IP phones, CAC, bandwidth control and management, and address translation. The call agent also serves as a repository for call routing information.

• Application servers Provide services such as voice mail, unified messaging, or call center support.

• Videoconference station Provides access for end-user participation in videoconferencing. The videoconference station contains a video capture device for video input and a microphone for audio input. The user can view video streams and hear the audio that originates at a remote user station. Cisco targets its VT Advantage product at desktop videoconferencing applications.

Other components, such as software voice applications, interactive voice response (IVR) systems, and softphones, provide additional services to meet the needs of enterprise sites.


Call Control

Call control allows users to establish, maintain, and disconnect a voice flow across a network, as shown in Figure 1-13.


Although different protocols address call control in different ways, they all provide a common set of services. The following are the basic components of call control:

• Call setup Checks call-routing configuration to determine the destination of a call. The configuration specifies the bandwidth requirements for the call. When the bandwidth requirements are known, CAC determines if sufficient bandwidth is available to support the call. If bandwidth is available, call setup generates a setup message and sends it to the destination. If bandwidth is not available, call setup notifies the initiator by presenting a busy signal. Different call control protocols, such as H.323, Media Gateway Control Protocol (MGCP), and Session Initiation Protocol (SIP), define different sets of messages to be exchanged during setup.

• Call maintenance Tracks packet count, packet loss, and interarrival jitter, or delay, when the call is set up. Information passes to the voice-enabled devices to determine if connection quality is good or if it has deteriorated to the point where the call should be dropped.

• Call teardown Notifies voice-enabled devices to free resources and make them available for the next call when either side terminates a call.

From a design perspective, you can set up call control in either a distributed or centralized architecture. The following sections describe both types.


Distributed Call Control

Distributed call control, an example of which is shown in Figure 1-13, offers an environment where call control is handled by multiple components in the network. This approach to call control is possible where the voice-capable device is configured to support call control directly. This is the case with a voice gateway when protocols, such as H.323 or SIP, are enabled on the device. In Figure 1-14, each location contains a Cisco Unified CallManager cluster. Each cluster is capable of handling call processing. Therefore, the topology shown demonstrates one example of distributed call control.


Distributed call control enables the gateway to perform the following procedure:

1. Recognize the request for service
2. Process dialed digits
3. Route the call
4. Supervise the call
5. Terminate the call


Centralized Call Control

Centralized call control, an example of which is illustrated in Figure 1-15, allows an external device (call agent) to handle the signaling and call processing, leaving the gateway to translate audio signals into voice packets after call setup. The call agent is responsible for all aspects of signaling, thus instructing the gateways to send specific signals at specific times. Also, the centralized call control model can leverage Cisco's Survivable Remote Site Telephony (SRST) feature to provide redundancy in the event of a WAN outage by having the voice-enabled router at the remote site perform basic call processing functions. In the figure, a Cisco Unified CallManager cluster located at the Headquarters location is in charge of call control. Therefore, the topology shown demonstrates an example of centralized call control.


When the call is set up, the following occur:

• The voice path runs directly between the two gateways and does not involve the call agent.

• When either side terminates the call, the call agent signals the gateways to release resources and wait for another call.

The use of centralized call control devices is beneficial in several ways:

• It centralizes the configuration for call routing and CAC. In a large voice environment, centralization can be extremely beneficial.

• The call agent is the only device that needs the intelligence to understand and participate in call control functions. These call control functions enable the customer to purchase less expensive voice-gateway devices and point to a single device to handle call control.

MGCP is one example of a centralized call control model.


Real-Time Versus Best-Effort Traffic

Voice and data can share the same medium. However, their traffic characteristics differ widely: Voice is real-time traffic and data is typically sent as best-effort traffic.

Traditional telephony networks were designed for real-time voice transmission, and therefore they cater to the need for a constant voice flow over the connection. Resources are reserved end to end on a per-call basis and are not released until the call is terminated. These resources guarantee that voice flows in an orderly manner. Good voice quality depends on the capacity of the network to deliver voice with guaranteed delay and timing.

Traditional data networks were designed for best-effort packet transmission. Packet telephony networks transmit with no guarantee of delivery, delay, or timing. Data handling is effective in this scenario because upper-layer protocols, such as TCP, provide for reliable, although untimely, packet transmission. TCP trades delay for reliability. Data can typically tolerate a certain amount of delay and is not affected by interpacket jitter.

A well-engineered, end-to-end network is required when converging delay-sensitive traffic, such as VoIP, with best-effort data traffic. Fine-tuning the network to adequately support VoIP involves a series of protocols and features to improve quality of service (QoS). Because the IP network is, by default, best effort, steps must be taken to ensure proper behavior of both the real-time and best-effort traffic. Packet telephony networks succeed, in large part, based on the QoS parameters that are implemented network-wide.