Troubleshooting Remote Access Networks (CCIE Professional Development)

Troubleshooting Remote Access

By Plamen Nedeltchev

Clocking, Line Coding, and Framing in Carrier Systems

In digital transmissions, three major elements need to be considered:

  • Clocking
  • Line coding (signaling)
  • Framing

They are the differentiators between analog and digital transmission.


Analog interfaces do not require that specific timing be configured. However, digital T1 interfaces require that not only the timing be set, but also that the origination of the timing source be considered. Line and internal timing (clocking) are the available options to address this requirement. Clocking refers to both timing and synchronization of the T1 and it is an essential part of the proper functioning. The timing is encoded in the frames and provides synchronization for the circuit and for every synchronous transmission. It is preferred to derive the clocking from the service provider, where the local exchange carrier (LEC) clocking serves as the master and the router clocking serves as a slave. When more than one provider is involved in delivering the T1, clock slips can occur. In this case, it is important to have a master clock to provide reliable time-division multiplexing (TDM) synchronization and alignment.

Pseudo-Ternary and Two Binary One Quaternary Signaling

ISDN and Frame Relay use a coding scheme called pseudo-ternary signaling, which is associated with the S/T interface.

The two binary one quaternary (2B1Q) coding scheme, which in turn is associated with the U-interface, is specific for U.S. ISDN, ISDN DSL (IDSL), high-data-rate DSL (HDSL), and single-line DSL (SDSL) local-loop specifications.

Pseudo-Ternary Line Coding and the S/T Interface

ISDN and the first layer of Frame Relay use a coding scheme called pseudo-ternary signaling, which is used by the S/T interface. This technique provides DC balance and non-voltage drifting procedures for signaling by using positive and negative 0s. In pseudo-signaling, 0 is represented as a line signal of approximately 750 millivolts that alternates between positive and negative polarity, and the 1 represents the absence of voltage. As you can see from Figure 2-6, synchronization is important because there must be a way to recognize the two consecutive 1s from 0s, where two or more consecutive 0s change the polarity. The synchronization is based on the bipolar violation (BPV), which is two consecutive 0s. In pseudo-ternary signaling, the binary values show a 0 as negative or positive, and 1 as the absence of signal.


Figure 2-6 Pseudo-Ternary Signaling

The frame structure is organized into blocks of bits. Figure 2-7 shows the short physical layer frame format. From the CPE terminal equipment (TE), the outgoing frames have 2 bits offset from incoming frames. The network timing drives the process. The TE cannot activate the line (activation bit A), as only the NT confirms the activation. Figure 2-7 shows a significant difference between incoming and outgoing frame formats. Each frame contains 16 bits from every B channel and 4 bits from the D channel.


Figure 2-7 Short and Detailed Frame Formats

The full description of the bits and rules follows:

  • Of— Offset.
  • F— Framing bit. Always positive 0 and is based on the BPV, which is used for timing.
  • L— DC balancing bit. Independent DC balanced; can be +0,-0, or 1.
  • D— D channel bit. +0, -0, or 1 in the first format; 1,-0 in the second.
  • E— Echo D channel bit. Can be +0, -0, or 1; exists only in the first format.
  • Fa Auxiliary framing bit. Only in the first format.
  • A— Activation bit. Can be +0, -0, or 1; only in the first format.
  • N— Complement to Fa. Can be +0, -0, or 1; only in the first format.
  • B1, B2— Data bit within B1 and B2 channels respectively; can be +0, -0, or 1.
  • S— Reserved.
  • M— Multiframing bit. Can be +0, -0, or 1; only in the first format. I.430 specifies multiframe, which is a group of 20 I.430 frames. M-bit and Fa are used in the multiframing procedures. The M-bit is set to 1 in the first 20 frames, which are 48 bits each, and set to 0 in every other frame. The I.430 also specifies that the first 0-bit transmitted F and L bits are another BPV.

For every 250 ms, 48 bits are transmitted through the channel, yielding a total of 4000 frames per second. 4000 frames x 48 bits each results in a total rate of 192 kb. For every 16 B-bits, the channel transmits 4 D-bits. Therefore in a Basic Rate Interface (BRI), if every B channel is 64 kbps, the D channel is 16 kbps. A BRI is also known as 2B+D. This signaling is not contention-free, so a simple contention-resolution mechanism was designed to prevent contentions. It is based on the following traffic considerations:

  • B channel traffic— No additional functionality is necessary to control access to the two B channels because each channel is dedicated to a given TE at any particular time.
  • Incoming D channel traffic— The D channel is available for use on all member TEs. The Link Access Procedure on the D channel (LAPD) addressing scheme (described later) is sufficient to resolve any unit and its destination, based on the fact that each LAPD frame includes an explicit address of every destination TE. All TEs can read this address and determine whether the frame was sent to them.
  • Outgoing D channel traffic— Access must be regulated, so that only one device transmits frames at a time.

The following is how the contention resolutions works:

  • When the TE is ready to transmit an LAPD frame, it listens to the stream of incoming D channel echo bits. If it detects a string of 1s of equal length to a threshold value Xi, where i = priority class for this LAPD frame, it can transmit. Otherwise, it waits because another TE is transmitting.
  • If several TEs transmit 0 at the same time, all are using the same polarity. Sending 1 means no signal. The TEs are attached to the bus in parallel. Based on Ohm's law, the total voltages are not a sum of all voltages. Therefore, 1 is detected if all TEs apply 1 (no voltage) and 0 is detected if one or more TEs apply a voltage. This process uses the logical AND function.
  • A NT-to-TE frame carries an E-bit, which is an echoed D-bit in the opposite direction. The E-bit performs an important contention mechanism, especially in the p2mp designs. The mechanism ensures that only one TE is transmitting frames in the TE-to-NT direction. If more than one TE tries to transmit, a collision can result. To avoid a collision, a transmitting TE monitors the echo bit with the transmitted bit. If the E-bit is different from the last transmitted D-bit by this TE, the TE knows that it does not control the D channel and ceases the transmission. This procedure is called perfect scheduling.

Two Binary One Quaternary Coding, the Frame Format, and the U-Interface

ANSI standard T1.601 is used in the U.S. to provide the necessary specifications for the U-reference point because the International Telecommunication Union Telecommunication Standardization Sector (ITU-T) does not define the local loop specifications between NT and the LE across U-reference points, or the otherwise called U-interface. The interface and line coding is specific for U.S. ISDN, IDSL, HDSL and SDSL local loop specifications. The physical connection over twisted pair provides distances up to 5.5 km (18,000 ft). The U-interface supports serial, synchronous, full-duplex, and point-to-point designs. The signaling technique associated with the U-interface is called two binary one quaternary (2B1Q), as shown in Figure 2-8.


Figure 2-8 Two Binary One Quaternary (2B1Q) Signaling

2B1Q is a four-level, single-symbol code. Every combination of two bits (first two columns), where the first bit represents the polarity and the second one the magnitude, has a voltage representation and Q symbol representation called a quat. The frame format in the U-access point (often called the U-interface) is different, as shown in Figure 2-9.


Figure 2-9 2B1Q Frame Format

The Synchronization Word (SW) is a self-explanatory term, which is used for physical layer synchronization and frame alignment. B and D channel information is formed into 12 groups that contain data and control information as follows: B1 (8 bits), B2 (8 bits), and D (2 bits). Therefore, 18 (8 + 8 + 2) multiplied by 12 yields 216 bits, representing 108 quats. The overhead field is used for channel maintenance, limited bit error detection, and power status indication. Using two different interfaces, S/T and U, requires signal transformation that is performed by the device known as the network termination type 1 or NT1. The timing for the NT1 in a U-interface (reference point) is still provided by the LE.

Unlike a S/T reference point, the U-access point operates at 160 kbps and sends 666.666 frames per second, or one frame every 1.5 ms. Every eight groups of frames is a SF. To indicate the beginning of the SF, the technique uses an inverted SW. The 6 overhead bits of all frames represent a 48-bit (8 frames x 6 bits) block, called a M-channel. This block is capable of increased error-detection, signaling maintenance, and power error detection.

T1 Digital Coding and Framing

Analog interfaces do not require a specific line-coding configuration. Digital interfaces do require that the alternate mark inversion (AMI), bipolar 8-zero substitution (B8ZS), B3ZS, B6ZS, or high density binary 3 (HDB3) be configured. These values must match the values of the private branch exchange (PBX) or central office (CO) that connects to the Digital T1/T3 Packet Voice Trunk Module.

A large variety of coding schemes can be used that are based on well-established research and standards. The following are a few of these coding schemes:

  • Unipolar non-return to zero (NRZ)
  • Unipolar return to zero
  • Polar NRZ
  • Polar NRZ inverted (NRZI)
  • Bipolar return to zero
  • B3ZS
  • B6ZS
  • Manchester code

Some of these are included in Figure 2-10.


Figure 2-10 Line-Coding Schemes

Choosing one or another scheme is always based on design considerations, such as the following:

  • Does the line code provide good synchronization?
  • Does the line code allow a DC buildup in transmission?
  • Does the line code provide any error detection capability?

Alternate Mark Inversion

Cisco routers support the following coding schemes:

7200-router(config-controller)#linecode ?
ami   AMI encoding
b8zs  B8ZS encoding

One of the most widespread coding techniques for T1 carriers is alternate mark inversion (AMI). By using AMI, pulses correspond to binary 1s and 0s and alternate at 3V (+3/-3V). The presence of a signal is 1, and the absence of a signal is 0. A benefit of this encoding is a built-in method of error detection. When consecutive pulses are detected to have the same polarity, the condition is considered to be a BPV. As a result, the carrier and the CPE indicate that the frame is experiencing some type of error. A problem with this coding scheme is the way that 0 is interpreted as a no-signal condition. Therefore, an absence of pulses (all 0s) can force repeaters and network equipment to lose frame synchronization. To prevent this, all T1s are required to meet the 1's density requirement, which states that no more than 15 consecutive 0s can be transmitted to the line. (In the old days, the FCC actually said you could have 15 0s in a row. Now, the FCC says that you can have up to about 40 0s without harming the network. For all practical purposes, 7 consecutive 0s is the maximum today. See Newton's Telecom Dictionary: The Authoritative Resource for Telecommunications, Networking, the Internet, and Information Technology (18th Edition), Harry Newton and Ray Horak.) One solution to this requirement is alternate space inversion (ASI), which reverts the pulses. 1 becomes a no-signal state and the data is inverted for high-level data link control (HDLC) framed packets. However, this isn't commonly used. More often, you find CPE equipment configured to only use 7 of the 8 bits of each T1 timeslot. This effectively brings your data rate down from 64 k to 56 k per channel.

Zero-Suppression Schemes

The most widespread remedies for the density requirements of T1s are zero-suppression codes. They use certain rules, where if a predefined pattern is detected, the network equipment corrects it by inserting another pattern in the data stream to maintain ones density.

B8ZS and T1

B8ZS is one of the most widespread zero-suppression coding techniques. It is implemented to prevent degradation because of long strings of 0s. B8ZS replaces a block of eight consecutive 0s with a code that contains BPVs in the fourth and seventh bits. When eight 0s appear, they are replaced with the B8ZS code before being multiplexed onto the T1 line. At the receiver, detection of the BPV is replaced with eight 0s, which allows the full 64 kbps of the DS0 to be used. This is the most common technique; all major U.S. carriers support it.

B3ZS and B6ZS for T3

Other well-known zero suppressing schemes are B3ZS and B6ZS line codes. These coding schemes are typical for T3 (digital service 3 [DS3]) circuits.

In B3ZS, each pattern of 000 is replaced by 00V or B0V. The choice depends on whether the number of bipolar pulses between violations (V) is an odd number. In this case, V is positive or negative and chosen to cause a BPV, and B is also positive or negative, and is chosen to meet the bipolar conditions.

In B6ZS, each pattern of 000000 is replaced by 0VB0VB. Again, the choice depends on whether the number of bipolar pulses between Vs is an odd number. Here, as in the previous case, V is positive or negative, and is chosen to cause a BPV; and B is positive or negative, and is chosen to meet the bipolar conditions.

T1 and T3 Framing

Analog interfaces do not require that specific framing be configured. Digital T1 interfaces do require that either SF (also called SF or D4 framing) or Extended SF (ESF) be configured. These values must match the values of the PBX or CO that connects to the Digital T1 Packet Voice Trunk Module.

Cisco routers support the following frame formats for T1 connections:

7200-router(config-controller)#framing ?
esf  Extended Superframe
sf   Superframe

T1 SF Signal Format

The recent D3/Mode 3 D4 format for framing and channelization is, by far, the most popular format. The bit stream is organized into SFs, each consisting of 12 frames. Every frame consists of channel information, where every channel is 8 bits plus framing bits. Framing bits are marked differently. Terminal framing bits (BFt) mark odd frames that produce a sequence of alternating 1s and 0s. Even-numbered signaling frame bits (BFs) produce groups of three 1s, followed by three 0s, repeating. The framing bits, which are every 193rd bit and the last bit in each frame, are inserted between the 24th and 1st channel word. Each channel word consists of 8 bits (B1 through B8) every 0.65 nsec. Channel words represent 8-bit samples, taken at the rate of 8000 samples per second, and correspond to 24 different sources of voice or data information.

Signaling information is information that is exchanged between components of a telecommunication system to establish, monitor, or release connections. For voice transmissions, signaling information must be transmitted with the channel voice samples. This is accomplished by sharing the last significant bit (B8) between voice and signaling. This process is called robbed bit signaling (RBS). The B8 bit carries voice information for five frames, followed by one frame of signaling information. This pattern of B8 assignment to voice and signaling is repeated during each successful group of six frames. Using this technique, 24 channels x (8000 samples/channel/second) x (8 bits per frame) + 8000 BFs/second, yields a total speed of 1.544 Mbps.

T1 ESF Signal Formats

The SF of ESF is extended from 12 to 24 frames with 24 framing bits. Of the 24 framing bits in an ESF, six bits are used for synchronization, six bits for error checking, and the remaining 12 bits are used for a 4-kbps facility data link (FDL), which is a communication link between CSUs and the telephone company's monitoring devices. The framing bits are used for different purposes than in SF. The ESF takes advantage of new more-reliable conduits (just an analogy with X.25 and Frame Relay standards), where not every bit needs to be used for framing and synchronization. To permit error detection, the sending CSU examines all the 4608 data bits within ESF and generates a cyclic redundancy check (CRC). The receiver calculates its own CRC and compares both. If there is a match, there are no errors. CRC is known to report approximately 98 percent of all possible bit errors. This information is stored in counters and considered when there is a trend in speed degradation. This process results in increased availability and uptime for T1s. Some reports show that the expected availability of T1s for most phone companies is greater than 95 percent.

M23 Frame Format

The digital signal level 3 (DS-3) (T3) interface operates at 44.736 Mbps over coax cable, which is compliant with Asynchronous Transfer Mode (ATM) Forum UNI specifications. Three standards for DS-3 framing exist: M23, C-bit parity, and SYNTRAN. The M23 multiplex scheme provides for the transmission of seven DS-2 channels. A T3 is 28 T1s, and the first layer multiplexor (M12) serves four T1s. The seven second-layer multiplexors are connected to the end multiplexor (M23). Because each DS-2 channel can contain four DS-1 signals, a total of 28 DS-1 signals (670 DS-0 signals) are transported in a DS-3 facility. The current DS-3 signal format is a result of a multi-step, partially synchronous, partially asynchronous multiplexing sequence. Cisco routers support the following frame formats:

7200-router(config)#contr t3 2/0
7200-router(config-controller)#framing ?
auto-detect  Application Identification Channel Signal
c-bit        C-Bit Parity Framing
m23          M23 Framing Format

The DS-3 signal is partitioned into M-frames of 4760 bits each. The M-frames are divided into seven M-subframes, each containing 680 bits. Each subframe is further divided into eight blocks of 85 bits each, with the first bit used for control and the rest for payload. 56 frame overhead bits handle functions such as M-frame alignment, M-subframe alignment, performance monitoring, alarms, and source application channels.

PRI—1.544-Mbps Interface

The ITU-T recommendation I.431 defines the physical layer protocol of PRI for both 1.544 and 2.048 Mbps. The electrical characteristics are defined in G.703 and G.704. The primary use of PRIs are as the trunk and trunk groups, and not as the TEs. Usually, LECs have separate service groups for BRIs and PRIs, which is based on the nature of their usage (see Figure 2-11).


Figure 2-11 ITU-T I.431 Recommendation

The PRI, unlike a BRI does not support p2mp configurations, but only point-to-point. In terms of ISDN, usually it is defined at the T reference point, where a digital PBX or LAN connection device controls multiple ISDN TEs, and provides multiplexing for them. The PRI is based on the DS1 transmission structure and T1 services.

The PRI multiplexes 24 channels, 64 kbps each. When it is configured as 23 B+D channels, the D channel is used for signaling, and when it is configured as 24 B channels, another D channel is available to do the signaling.

When the PRI is configured for 24 B channels, the PRI frame contains one framing bit plus a single 8-bit pulse code modulation (PCM) sample from each of the 24 channels. These 193 bits, multiplied by 8000 frames per second, yield a total bit rate of 1.544 Mbps. The 24 frames are grouped together to form a multiframe, as described in the ESF frame format. One multiframe is 24 bits and performs the following functions:

  • Frame alignment sequence (FAS)— FAS ensures the synchronization of frames and is represented by a repeated pattern of 001011.... In case of losing proper synchronization, the receiver needs to listen to five consecutive SFs and find the correct sequence to synchronize.
  • Flow Sequence Control (FSC)— FSC ensures that no bit errors occurred in the previous multiframe. The sequence detects and reports, but does not correct the errors.
  • Maintenance channel (M)— The use of M remains to be determined; it is not currently used in PRIs.

In I.431, one bit of every octet is required to meet the density requirement, resulting in a 56-kbps channel rate. To overcome this drawback, B8ZS is used and all 0 octets are replaced by a combination of 000110011, where the BPV occurs in the fourth and seventh bits. If the last 1 was a positive 1, the replacement looks like 000pn0np, where n and p represent negative and positive polarity. If the previous appearance of 1 was negative, the combination is 000np0pn, and n and p are negative and positive polarities. The new pattern is injected in a way that all V and B are paired, so that the new code is DC-balanced.

PRI—2.048-Mbps Interface

Similar to the 1.544-Mbps PRI used in the U.S. and based on the T1 carrier, the 2.048-Mbps PRI is typical for Europe and is based on Computer Emergency Response Team (CERT) 1, or the E1 carrier. E1, in the configuration of 30B+D, is the most common remote access solution for ISDN access services. Thirty B-channels and configurations such as 30B+D are more likely to be used for home remote access solutions. The numeration of the channels differs from the previous design (see Figure 2-11). Channels 1 to 15 are B channels, followed by a D channel (number 16), and another set of 15 B channels. Every frame contains 31 slots, and every slot is a single 8-bit sample. The duration of the frame is still 125 seconds, but 256 bits. So, 8000 frames per second with 256 bits each, plus framing bits, yield 2.048 Mbps with 1.9984 Mbps for the user data rate. Unlike AMI in the previous design, the zero-suppression scheme is called HDB3. HDB3 replaces a string of four consecutive 0s with a pattern of x00V, where V is the violation bit, and x is the bit which can be 1 or 0, depending on the requirement to keep the code DC-balanced.

The line-coding and framing techniques described previously include the most common and established ones, but not all of them. The main objective is to provide you with background and details to prepare for the next chapter.

+ Share This