SONET and SDH: Advanced Topics

This chapter is from the book

Optical Network Control: Architecture, Protocols, and Standards

Learn More Buy

This chapter is from the book

This chapter is from the book 

Optical Network Control: Architecture, Protocols, and Standards

Learn More Buy

3.6 When Things Go Wrong

One of the most important aspects built into optical transport systems is their “self-diagnosis” capability. That is, the ability to detect a problem (i.e., observe a symptom), localize the problem (i.e., find where it originated), and discover the root cause of the problem. In fact, SONET and SDH include many mechanisms to almost immediately classify the root cause of problem. This is done by monitoring the signal integrity between peers at a given layer, and also when transferring a signal from a client (higher) layer into a server (lower) layer (Figure 2-17).

In the following, we first consider the various causes of transport problems. Next, we examine how problems are localized and how signal quality is monitored. Finally, we review the methods and terminology for characterizing problems and their duration.

3.6.1 Transport Problems and Their Detection

Signal monitoring functionality includes the following: continuity supervision, connectivity supervision, and signal quality supervision. These are described next.

3.6.1.1 CONTINUITY SUPERVISION

A fundamental issue in telecommunication is ascertaining whether a signal being transmitted is successfully received. Lack of continuity at the optical or electrical layers in SONET/SDH is indicated by the Loss of Signal (LOS) condition. This may arise from either the failure of a transmitter (e.g., laser, line card, etc.) or break in the line (e.g., fiber cut, WDM failure, etc.). The exact criteria for when the LOS condition is declared and when it is cleared are described in reference [ITU-T00b]. For optical SDH signals, a typical criterion is the detection of no transitions on the incoming signal (before unscrambling) for time T, where 2.3 µs ≤ T ≤ 100 µs. An LOS defect is cleared if there are signal transitions within 125 µs. When dealing with other layers, the loss of continuity is discovered using a maintenance signal known as the Alarm Indication Signal (AIS). AIS indicates that there is a failure further upstream in the lower layer signal. This is described further in section 3.6.2.1.

3.6.1.2 CONNECTIVITY SUPERVISION

Connectivity supervision deals with the determination of whether a SONET/SDH connection at a certain layer has been established between the intended pair of peers. This is particularly of interest if there has been an outage and some type of protection or restoration action has been taken. A trail trace identifier is used for connection supervision. Specifically,

• The J0 byte is used in the SONET section (SDH regenerator section) layer. The section trace string is 16 bytes long (carried in successive J0 bytes) as per recommendation G.707 [ITU-T00a].

• The J1 byte is used in the SONET/SDH higher-order path layer (e.g., SONET STS-1 and above). The higher-order path trace string could be 16 or 64 bytes long as per recommendation G.707 [ITU-T00a].

• The J2 byte is used in the SONET/SDH lower-order path layer (e.g., SONET VT signals). The lower-order path trace string is 16 bytes long as per recommendation G.707 [ITU-T00a].

For details of trail trace identifiers used for tandem connection monitoring (TCM), see recommendations G.707 [ITU-T00a] and G.806 [ITU-T00c]. The usage of this string is typically controlled from the management system. Specifically, a trace string is configured in the equipment at the originating end. An “expected string” is configured at the receiving end. The transmitter keeps sending the trace string in the appropriate overhead byte. If the receiver does not receive the expected string, it raises an alarm, and further troubleshooting is initiated.

3.6.1.3 SIGNAL QUALITY SUPERVISION

Signal quality supervision determines whether a received signal contains too many errors and whether the trend in errors is getting worse. In SONET and SDH, parity bits called Bit Interleaved Parity (BIP) are added to the signal in various layers. This allows the receiving end, known as the near-end, to obtain error statistics as described in section 3.6.3. To give a complete view of the quality of the signal in both directions of a bidirectional line, the number of detected errors at the far-end (transmitting end) may be sent back to the near-end via a Remote Error Indicator (REI) signal.

The following bits and bytes are used for near-end signal quality monitoring under SONET and SDH:

• SONET section (SDH regenerator section) layer: The B1 byte is used to implement a BIP-8 error detecting code that covers the previous frame.

• SONET line (SDH multiplex section) layer: In the case of SDH STM-N signals, a BIP N × 24 composed of the 3 STM-1 B2 bytes is used. In the case of SONET STS-N, a BIP N × 8 composed of the N B2 bytes is used. These cover the entire contents of the frame excluding the regenerator section overhead.

• SONET path (SDH HOVC) layer: The B3 byte is used to implement a BIP-8 code covering all the bits in the previous VC-3, VC-4, and VC-4-Xc.

• SONET VT path (SDH LOVC) layer. Bits 1 and 2 of the V5 byte are used to implement a BIP-2 code covering all the bits in the previous VC-1/2.

SONET/SDH provides the following mechanisms for carrying the REI information. For precise usage, see either T1.105 [ANSI-95a] or G.707 [ITU-T00a].

• Multiplex section layer REI: For STM-N (N = 0, 1, 4, 16), 1 byte (M1) is allocated for use as Multiplex Section REI. For STM-N (N = 64 and 256), 2 bytes (M0, M1) are allocated for use as a multiplex section REI. Note that this is in line with the most recent version of G.707 [ITU-T00a].

• Path layer REI: For STS (VC-3/4) path status, the first 4 bits of the G1 path overhead are used to return the count of errors detected via the path BIP-8, B3. Bit 3 of V5 is the VT Path (VC-1/2) REI that is sent back to the originating VT PTE, if one or more errors were detected by the BIP-2.

3.6.1.4 ALIGNMENT MONITORING

When receiving a time division multiplexed (TDM) signal, whether it is electrical or optical, a critically important stage of processing is to find the start of the TDM frame and to maintain frame alignment. In addition, when signals are multiplexed together under SONET/SDH, the pointer mechanism needs to be monitored.

Frame Alignment and Loss of Frame (LOF)

The start of an STM-N (OC-3N) frame is found by searching for the A1 and A2 bytes contained in the STM-N (OC-3N) signal. Recall that the A1 and A2 bytes form a particular pattern and that the rest of the frame is scrambled. This framing pattern is continuously monitored against the assumed start of the frame. Generally, the receiver has 625 µs to detect an out-of-frame (OOF) condition. If the OOF state exits for 3 ms or more then a loss of frame (LOF) state will be declared. To exit the LOF state, the start of the frame must be found and remain valid for 3 ms.

Loss of Multiframe

SDH LOVCs and SONET VTs use the multi-frame structure described earlier. The 500 µs multiframe start phase is recovered by performing multiframe alignment on bits 7 and 8 of byte H4. Out-of-multiframe (OOM) is assumed once when an error is detected in the H4 bit 7 and 8 sequence. Multiframe alignment is considered recovered when an error-free H4 sequence is found in four consecutive VC-n (VT) frames.

Pointer Processing and Loss of Pointer (LOP)

Pointer processing in SONET/SDH is used in both the HOVC (STS path) and LOVC (VT path) layers. This processing is important in aligning payload signals (SDH VC or SONET paths) into their containing signals (STM-N/OC-3N). Without correct pointer processing, essentially one per payload signal, the payload signal is essentially “lost.” Hence, pointer values are closely monitored as part of pointer processing [ITU-T00a, ITU-T00b]. A loss of pointer state is declared under severe error conditions.

3.6.2 Problem Localization and Signal Maintenance

Once a problem has been detected, its exact location has to be identified for the purposes of debugging and repair. SONET/SDH provides sophisticated mechanisms to this in the form of Alarm Indication Signals (AIS) and the Remote Defect Indication (RDI). These are described below.

3.6.2.1 ALARM INDICATION SIGNALS

Suppose that there is a major problem with the signal received by an intermediate point in a SONET network. In this case, a special Alarm Indication Signal is transmitted in lieu of the normal signal to maintain transmission continuity. An AIS indicates to the receiving equipment that there is a transmission interruption located at, or upstream, of the equipment originating the AIS. Note that if the AIS is followed upstream starting from the receiver, it will lead to the location of the error. In other words, the AIS signal is an important aid in fault localization. It is also used to deliver news of defects or faults across layers.

A SONET STE will originate an Alarm Indication Signal-Line (AIS-L) (MS AIS in SDH) upon detection of an LOS or LOF defect. There are two variants of the AIS-L signal. The simplest is a valid section overhead followed by “all ones” pattern in the rest of the frame bytes (before scrambling). To detect AIS-L, it is sufficient to look at bits 6, 7, and 8 of the K2 byte and check for the “111” pattern. A second function of the AIS-L is to provide a signal suitable for normal clock recovery at downstream STEs and LTEs. See [ANSI95a] for the details of the application, removal, and detection of AIS-L.

A SONET LTE will generate an Alarm Indication signal-Path (AIS-P) upon detection of an LOS, LOF, AIS-L, or LOP-P defect. AIS-P (AU AIS in SDH) is specified as “all ones” in the STS SPE as well as the H1, H2, and H3 bytes. STS pointer processors detect AIS-P as “111...” in bytes H1 and H2 in three consecutive frames.

A SONET STS PTE will generate an Alarm Indication signal-VT (AIS-V) for VTs of the affected STS path upon detection of an LOS, LOF, AIS-L, LOP-P, AIS-P, or LOP-V defect. The AIS-V signal is specified as “all ones” in the entire VT, including the V1-V4 bytes. VT pointer processors detect AIS-V as “111...” in bytes V1 and V2 in three consecutive VT superframes.

The SDH AIS signals for its various layers are nearly identical as those of SONET in definition and use as shown in Table 3-5.

3.6.2.2 REMOTE DEFECT INDICATION

Through the AIS mechanism, SONET allows the downstream entities to be informed about problems upstream in a timely fashion (in the order of milliseconds). The AIS signal is good for triggering downstream protection or restoration actions. For quick recovery from faults, it is also important to let the upstream node know that there is a reception problem downstream. The Remote Defect Indication (RDI) signal is used for this purpose. The precise definition of RDI, as per [ANSI95a], is

Table 3-5. SDH AIS Signals by Layer

A signal transmitted at the first opportunity in the outgoing direction when a terminal detects specific defects in the incoming signal.

At the line level, the RDI-L code is returned to the transmitting LTE when the receiving LTE has detected an incoming line defect. RDI-L is generated within 100 ms by an LTE upon detection of an LOS, LOF, or AIS-L defect. RDI-L is indicated by a 110 code in bits 6,7,8 of the K2 byte (after unscrambling).

At the STS path level, the RDI-P code is returned to the transmitting PTE when the receiving PTE has detected an incoming STS path defect. There are three classes of defects that trigger RDI-P:

1. Payload defects: These generally indicate problems detected in adapting the payload being extracted from the STS path layer.

2. Server defects: These indicate problems in one of the layers responsible for transporting the STS path.

3. Connectivity defects: This only includes the trace identifier mismatch (TIM) or unequipped conditions.

Table 3-6 shows current use of the G1 byte for RDI-P purposes (consult [ANSI95a] for details).

The remote defect indication for the VT path layer, RDI-V, is similar to RDI-P. It is used to return an indication to the transmitting VT PTE that the receiving VT PTE has detected an incoming VT Path defect. There are three classes of defects that trigger RDI-V:

Table 3-6. Remote Defect Indicator—Path (RDI-P) via the G1 Byte

1. Payload defects: These generally indicate problems detected in adapting the payload being extracted from the VT path layer.

2. Server defects: These generally indicate problems in the server layers to the VT path layer.

3. Connectivity defects: These generally indicate that there is a connectivity problem within the VT path layer.

For more information, see [ANSI95a] for details. RDI-V uses the Z7 bytes (bits 6 and 7).

One thing to note about RDI signals is that they are “peer to peer” indications, that is, they stay within the layer that they are generated. The AIS and RDI signals form the “fast” notification mechanisms for protection and restoration, that is, these are the primary triggers. Examples of their usage are given in the next chapter. The RDI signals in various SDH layers are nearly identical to those of SONET and they are summarized in Table 3-7.

Table 3-7. RDI Signals for Various SDH Layers

3.6.3 Quality Monitoring

3.6.3.1 BLIPS AND BIPS

The bit error rates are typically extremely low in optical networks. For example, in 1995, the assumed worst-case bit error rate (BER) for SONET regenerator section engineering was 10^-10, or one error per 10 billion bits. Today, that would be considered quite high. Hence, for error detection in a SONET frame, we can assume very few bit errors per frame.

As an example, the number of bits in an STS-192 frame is 1,244,160 (9 rows × 90 columns per STS-1 x 8 bits/byte × 192 STS-1). With a BER of 10^-10, it can be expected that there will be one bit error in every 8038 frames. The probability of two errors in the same frame is fairly low. Since the bit rate of an STS-192 signal is 10 Gbps (or 10¹⁰ bits per second), a BER of 10^-10 gives rise to one bit error every second on the average. This is why a BER of 10^-10 is considered quite high today.

Figure 3-10 shows the general technique used in SONET and SDH for monitoring bit errors “in-service” over various portions of the signal. This method is known as the Bit Interleaved Parity 8 Bits, or BIP-8 for short. Although the name sounds complex, the idea and calculation are rather simple. In Figure 3-10, X1-X5 represents a set of bytes that are being checked for transmission errors. For every bit position in these bytes, a separate running tally of the parity (i.e., the number of 1s that occur) is kept track of. The corresponding bit position of the BIP-8 byte is set to “1” if the parity is currently odd and a zero if the parity is even. The BIP-8 byte is sent, typically in the following frame, to the destination. The destination recomputes the BIP-8 code based on the contents of the received frame and compares it with the BIP-8 received. If there are no bit errors, then these two codes should match. Figure 3-10(b) depicts the case where one of the bytes, X2, encounters a single bit error during transmission, that is, bit 2 changes from 1 to 0. In this case, the received BIP-8 and the recomputed BIP-8 differ by a single bit and, in fact, the number of differing bits can be used as an estimate of the number of bit errors.

Figure 3-10. Example of BIP-8 Calculation and Error Detection

Note that the BIP-8 technique works well under the assumption of low bit error rates. The study of general mechanisms for error detection and correction using redundant information bits is known as algebraic coding theory (see [Lin+83]).

BIP-8 is used for error monitoring in different SONET/SDH layers. At the SONET section layer, the B1 byte contains the BIP-8 calculated over all the bits of the previous STS-N frame (after scrambling). The computed BIP-8 is placed in the B1 byte of the first STS-1 (before scrambling). This byte is defined only for the first STS-1 of an STS-N signal. SDH uses this byte for the same purpose. Hence, the BIP-8 in this case is calculated over the entire SONET frame and covers a different number of bytes for different signals, for example, STS-12 vs. STS-192.

At the SONET line layer, BIP-8 is calculated over all the bits of the line overhead and the STS-1 SPE (before scrambling). The computed BIP-8 is placed in the B2 byte of the next STS-1 frame (before scrambling). This byte is separately computed for all the STS-1 signals within an STS-N signal. These N BIP-8 bytes are capable of detecting fairly high bit error rates, up to 10^-3. To see this, consider an STS-1 line signal (i.e., an STS-1 frame without section layer overhead). The number of bytes in this signal is 804 (9 rows × 90 columns – 6 section bytes). Each bit in the line BIP-8 code is used to cover 804 bits (which are in the corresponding bit position of the 804 bytes in the line signal). Since a BER of 10^-3 means an average of one bit error every 1000 bits, there will be less than one bit error in 804 bits (on the average). This, the line BIP-8 code is sufficient for detecting these errors. Note, however, that BIP-8 (and any parity based error detection mechanism) may fail if there are multiple, simultaneous bit errors.

At the STS Path level, BIP-8 is calculated over all the bits of the previous STS SPE (before scrambling) and carried in the B3 path overhead byte. SDH uses this byte for the same purpose but excludes the fixed stuff bytes in the calculation. The path BIP-8, like the section BIP-8, covers a different number of bytes depending on the size of the STS path signal, that is, STS-3 vs. STS-12.

At the VT path level, 2 bits of the VT path level overhead byte V5 are used for carrying a BIP-2. The technique for this is illustrated in Figure 3-11. To save on overhead, the parity counts over all the odd and the even bit positions are combined and represented by the two bits of the BIP-2 code, respectively. Recall that the VT SPE is a multiframe spanning four SONET frames. The BIP-2 is calculated over all bytes in the previous VT SPE, including all overhead but the pointers (Figure 3-11).

Figure 3-11. BIP Calculation at the VT Path Level

Let us examine how effective the BIP-2 code is. The number of bits in the VT1.5 SPE is 832 ([(9 rows × 3 columns) – 1 pointer byte] × 8 bits/byte × 4 frames per SPE). Each bit of the BIP-2 code covers half the bits in the VT1.5 SPE, that is, 416 bits. Hence, BIP-2 can handle error rates of 1 in 500 bits (BER between 10^-2 and 10^-3). Now, a VT6 is four times the size of the VT1.5. In this case, each parity bit covers 1664 bits, handling a BER slightly worse than 10^-4.

3.6.4 Remote Error Monitoring

The error monitoring capabilities provided by SONET and SDH enables the receiver to know the error count and compute the BER on the received signal at various layers. Based on this information, it is useful to let the sender learn about the quality of the signal received at the other end. The following mechanisms are used for this purpose.

The STS-1 line REI (M0 byte) is used by the receiver to return the number of errored bits detected at the line layer to the sender. The receiver arrives at this number by considering the difference between the received and the recomputed BIP-8 (B2) codes. In the case of an STS-N signal, the M1 byte is used for conveying the REI information. Clearly, up to 8 × N errors could be detected with STS-N BIP-8 codes (as each STS-1 is covered by its own BIP-8). But only a count of at most 255 can be reported in the single M1 byte. Thus, in signals of OC-48 and higher rates, the number 255 is returned when 255 or more errors are detected.

At the path layer, the receiver uses the first four bits of the G1 path overhead to return the number of errors detected (using the path BIP-8) to the sender. At the VT path layer, the receiver uses bit 3 of the V5 byte to indicate the detection of one or more errors to the sender.

3.6.5 Performance Measures

When receiving word of a problem, one is inclined to ask some general questions such as: “How bad is it? “How long has it been this way?” and “Is it getting worse or better?” The following terminology is used in the transport world. An anomaly is a condition that gives the first hint of possible trouble. A defect is an affirmation that something has indeed gone wrong. A failure is a state where something has truly gone wrong. Whether an event notification or an alarm is sent to a management system under these conditions is a separate matter. Performance parameters in SONET and SDH are used to quantify these conditions.

A SONET or an SDH network element supports performance monitoring (PM) according to the layer of functionality it provides. A SONET network element accumulates PM data based on overhead bits at the Section, Line, STS Path, and VT Path layers. In addition, PM data are available at the SONET Physical layer using physical parameters. The following is a summary of different performance parameter defined in SONET. Similar performance parameters are also monitored and measured in SDH. For a detailed treatment on PM parameters on SONET refer to [Telcordia00].

Physical Layer Performance Parameters

The physical layer performance measurement enables proactive monitoring of the physical devices to facilitate early indication of a problem before a failure occurs. Several physical parameters are measured, including laser bias current, optical power output by the transmitter, and optical power at the receiver. Another important physical layer parameter is the Loss of Signal (LOS) second, which is the count of 1-second intervals containing one or more LOS defects^.

Section Layer Performance Parameters

The following section layer performance parameters are defined in SONET. Note that all section layer performance parameters are defined for the near-end. There are no far-end parameters at the Section layer.

• Code Violation (CV-S): The CV-S parameter is a count of BIP errors detected at the section layer. Up to eight section BIP errors can be detected per STS-N frame.

• Errored Second (ES-S): The ES-S parameter is a count of the number of 1 second intervals during which at least one section layer BIP error was detected, or an SEF (see below) or LOS defect was present.

• Errored Second Type A (ESA-S) and Type B (ESB-S): ESA-S is the count of 1-second intervals containing one CV-S, and no SEF or LOS defects. ESB-S is the count of 1-second intervals containing more than one but less than X CV-S errors, and no SEF or LOS defects. Here, X is a user-defined number.

• Severely Errored Second (SES-S): The SES-S parameter is a count of 1-second intervals during which K or more Section layer BIP errors were detected, or an SEF or LOS defect was present. K depends on the line rate and can be set by the user.

• Severely Errored Frame Second (SEFS-S): The SEFS-S parameter is a count of 1-second intervals during which an SEF defect was present. An SEF defect is detected when the incoming signal has a minimum of four consecutive errored frame patterns. An SEF defect is expected to be present when an LOS or LOF defect is present. But there may be situations when this is not the case, and the SEFS-S parameter is only incremented based on the presence of the SEF defect.

Line Layer Performance Parameters

At the SONET line layer, both near-end and far-end parameters are monitored and measured. Far-end line layer performance is conveyed back to the near-end LTE via the K2 byte (RDI-L) and the M0 or M1 byte (REI-L). Some of the important near-end performance parameters are defined below. The far-end parameters are defined in a similar fashion.

• Code Violation (CV-L): The CV-L parameter is a count of BIP errors detected at the line layer. Up to 8N BIP errors can be detected per STS-N frame.

• Errored Second (ES-L): The ES-L parameter is a count of 1-second intervals during which at least one line layer BIP error was detected or an AIS-L defect is present.

• Errored Second Type A (ESA-L) and Type B (ESB-L): ESA-L is the count of 1-second intervals containing one CV-L error and no AIS-L defects. ESB-L is the count of 1-second intervals containing X or more CV-L errors, or one or more AIS-L defects. Here, X is a user-defined number.

• Severely Errored Second (SES-L): The SES-L parameter is a count of 1-second intervals during which K or more line layer BIP errors were detected, or an AIS-L defect is present. K depends on the line rate and can be set by the user.

• Unavailable Second (UAS-L): Count of 1-second intervals during which the SONET line is unavailable. The line is considered unavailable after the occurrence of 10 SES-Ls.

• AIS Second (AISS-L): Count of 1-second intervals containing one or more AIS-L defects.

Path Layer Performance Parameters

Both STS path and VT path performance parameters are monitored at the path layer. Also, both near-end and far-end performance parameters are measured. Far-end STS path layer performance is conveyed back to the near-end STS PTE using bits 1 through 4 (REI-P) and 5 through 7 (RDI-P) of the G1 byte. Far-end VT path layer performance is conveyed back to the near-end VT PTE using bit 3 of the V5 byte (REI-V), and either bits 5 through 7 of the Z7 byte or bit 8 of the V5 byte (RDI-V). Some of the important near-end STS path performance parameters are defined below. The far-end parameters are defined in a similar fashion.

• Code Violation (CV-P): Count of BIP-8 errors that are detected at the STS-path layer.

• Errored Second (ES-P): Count of 1-second intervals containing one or more CV-P errors, one or more AIS-P, LOP-P, TIM-P, or UNEQ-P defects.

• Errored Second Type A (ESA-P) and Type B (ESB-P): ESA-P is the count of 1-second intervals containing one CV-P error and no AIS-P, LOP-P, TIM-P, or UNEQ-P defects. ESB-P is the count of 1-second intervals containing more than one but less than X CV-P errors and no AIS-P, LOP-P, TIM-P, or UNEQ-P defects. Here, X is a user-defined number.

• Severely Errored Second (SES-P): Count of 1-second intervals containing X or more CV-P errors, one or more AIS-P, LOP-P, TIM-P, or UNEQ-P defects. Here, X is a user-defined number.

• Unavailable Second (UAS-P): Count of 1-second intervals during which the SONET STS-path is unavailable. A path is considered unavailable after the occurrence of 10 SESs.

• Pointer Justification Counts: To monitor the adaptation of the path payloads into the SONET line, the pointer positive and negative adjustment events are counted. The number of 1-second intervals during which a pointer adjustment event occurs is also kept track of.