2.4. SONET/SDH Layering

Analogous to the layering models seen in packet networks [Stallings03], the TDM optical networks can also be layered according to the functions performed. Figure 2-17 shows layer diagrams of SONET and SDH, with the overhead usage and functionality illustrated. The VT path (lower-order VC) and the STS path (higher-order VC) level signals were previously described as being the envelopes containing the end-to-end user information. The physical layer is used to map the SONET frames into standardized optical signals. Each layer transports information from the layer immediately above it transparently, that is, without changing the content while adding some type of transport networking functionality.

2.4.1. The SONET Section (SDH Regenerator Section) Layer

Optical signals need to be regenerated after they have traversed a certain distance, which depends on a number of physical factors. This process can be applied to the signal amplitude (regenerate the signal), the signal's shape (reshaping), and timing (retiming). This is sometimes called 3R regeneration, and this happens at the physical layer. The SONET section layer is concerned with framing (A1 and A2 bytes) and scrambling. Since both these functions are key to re-timing, the section layer is closely coupled to the physical layer. The section layer, however, adds significant functionality to these basic regenerator functions in the form of error monitoring (B1 byte), a data communications channel (D1-D3 bytes), a signal trace (J0 byte), a user channel (F1 byte), and a local order-wire (E1) (see Figure 2-5).

These functions were incorporated into SONET to support regenerator equipment. Thus, the SDH terminology for this is regenerator section layer. As an example, suppose there is a long-haul fiber optic link, which requires the use of 3R regenerators. Then, the SONET section layer allows the performance (Bit Error Rate—BER) to be monitored on the segments between successive regenerators; the data communication channel permits communications with regenerator equipment for control and management purposes (see Chapter 6); the section trace (J0) feature aids in neighbor discovery and port connectivity verification (see Chapter 6); and the local order-wire can be used to furnish basic telephony communications at regenerator equipment sites. Equipment that terminates SONET section layer overhead is called Section Terminating Equipment (STE).

2.4.1.1. SCRAMBLING

Given an STS-N (or the corresponding SDH STM signal), all the A1, A2, J0/Z0 bytes are transported without modification. All the other bytes, both payload and overhead, are subject to “scrambling.” The framing bytes are sent without being scrambled so that the beginning of the frame can be easily delimited. Scrambling the rest of the frame prevents the unintentional introduction of the framing pattern of N A1s and A2s anywhere else into the frame (which could lead to the receiver becoming misaligned). Misalignment is called a Loss of Frame (LOF) condition, and it results in the loss of all data within the SONET/SDH signal. Another reason for scrambling is to maintain an adequate number of bit transitions. This enables clock recovery at the receiver. Inability to recover the signal timing at the receiver is one cause of the Loss of Signal (LOS) condition, which results in all the data in the signal being lost.

At SONET rates, scrambling is done in hardware with a simple circuit that (bit-wise) adds a pseudorandom bit sequence to all the bits in the frame (with the exceptions noted above). At the receiving end, the same pseudo random sequence is (bit-wise) “subtracted” from the signal. The pseudorandom sequence has nice “randomness” properties, but it actually repeats (if it did not repeat, the transmitter's and the receiver's scramblers cannot be synchronized). These types of sequences and the circuits to produce them are typically studied in texts on spread-spectrum communication [Holmes82, Sklar88]. In SONET/SDH, the scrambler's bit sequence repeats every 127 bits and is restarted at the beginning of every frame.

We can determine the number of payload bytes that pass until the scrambler sequence repeats. This is in fact 16 bytes (┌127 bits / 8 bits per byte┐ = 16). Is it possible to inject a false framing pattern in a VT structured STS-1? Due to the byte interleaving of VTGs and VTs in an STS-1, this would only be possible with synchronized coordination across multiple VTs and VTGs. Thus, this is highly unlikely. Is it possible to inject (willfully and maliciously) a false framing pattern in an STS-Nc carrying a data payload? Unfortunately, in early versions of Packet over SONET, this was possible. We will see how this problem was addressed in the next chapter.

One method of generating a pseudorandom bit sequence is by using a sequence of shift registers and binary adders (XOR gates). Such a circuit to generate the section layer scrambling sequence is shown in Figure 2-18. The bits (the D flip-flops shown) are reset to 1 at the beginning of the SONET/SDH frame. With every clock tick they are shifted right one position. Assuming the input data stream is all zeros, 256 bits of the output stream are:

     11111110000001000001100001010001111001000101100111010100111110
     10000111000100100110110101101111011000110100101110111001100101
     01011111110000001000001100001010001111001000101100111010100111
     11010000111000100100110110101101111011000110100101110111001100
     10101011.

It can be verified that this sequence is 127 bits long, and it repeats (beginning with 1111111).

2.4.2. The SONET Line (SDH Multiplex Section) Layer

The SONET line (SDH multiplex section) layer provides error monitoring (B2 byte) capability, a data communications channel (bytes D4-D12) with thrice the capacity of the section layer DCC, and an order-wire (E2 byte) capability (see Figure 2-5). The main purpose of the SONET line layer, however, is to provide multiplexing and synchronization support for STS path signals (SDH VC-3s and VC-4s). The pointer bytes (H1, H2, and H3) in the line overhead specify the location of the STS path signals within the payload and allow frequency compensation. Note that with concatenation, all the STS path signals contained within an STS-N do not need to be of the same size. For example, an STS-192 (OC-192) signal can support STS-1, STS-3c, STS-12c, and STS-48c paths all at the same time (depending on time slot availability). Because multiplexing is the key purpose of the SONET line layer, SDH chose to give this layer a bit more descriptive name, that is, the multiplex section layer.

In addition to the above essential capabilities, the line layer adds support for remote error indication (M0 byte), synchronization status messaging (S1 byte), and line automatic protection switching (APS) via the K1 and K2 bytes. Error monitoring, remote error indications, and alarms are discussed in the next chapter. Synchronization in SONET/SDH networks was covered in section 2.3.3.3, and APS is covered in Chapter 4.

2.4.3. The Tandem Connection Layer

The tandem connection layer is a very thin optional sublayer that allows the network, as opposed to the Path Termination Equipment (PTE), to monitor the health of one or more STS path signals, including concatenated signals [ANSI94a]. The N1 byte of the STS SPE (Figure 2-7) is allocated for this purpose. Four bits in this byte are used for error monitoring and the others are used to form a 32 Kbps path data channel. Although they share the same byte, the path data channel can be used independent of the bits being used for error monitoring.

2.4.4. The SONET Path (SDH Higher-Order Virtual Container) Layer

The SONET path layer is responsible for mapping payloads into a format convenient for multiplexing at the line layer. To this end, it provides some of the basic functionality of signal trace (J1 byte), error monitoring (B3) byte, and a user channel.

Since various signals can be mapped into a path layer payload, a byte of overhead is dedicated to the path signal label (byte C2). Table 2-3 gives a list of currently defined mappings.

Note that two of the most important protocols are not listed in Table 2-3: IP and Ethernet. As we will see in the next chapter, there do exist IP over SONET standards, and emerging standards for Ethernet and a number of other data communications protocols over SONET and SDH. The extended use of the signal label in these cases can ease interoperability and debugging, and could permit automatic configuration of PTE capable of processing multiple data protocols.

The frame structure of the STS-1 path signal was shown in Figure 2-7. In addition to error monitoring via the B3 byte, the path layer provides a mechanism for the terminating PTE to report the path termination status and performance (error rates) to the originating PTE. The path status byte G1 is used for this purpose.

When VT signals are multiplexed into an STS path signal, two bits from the multiframe indicator byte H4 are used to count from 0 to 3 repeatedly for each 125 µs SONET frame. This 500 µs repeating structure is the superframe discussed earlier. When the multiframe indicator is used in combination with VT pointers or overhead, it turns the one allocated byte of VT pointer per frame into 4 bytes of VT pointer per superframe.

2.4.5. The VT Path Layer

Before we discuss the VT path layer, we have to examine the VT pointer bytes V1 through V4. These occur as the first byte of the VT structures, as shown in Figures 2-10–2-12. The VT structures repeat four times within the superframe. Of these bytes, only V1–V3 are actually used. Also, all but 2 bits of the combined V1 and V2 bytes are dedicated to pointer values and operations, and the remaining 2 bits indicate the size of the VT.

How does the SONET VT processing hardware locate VTs within an STS SPE if the seven VTGs that comprise that SPE contain different types of VTs? In other words, how does the hardware know where the pointer bytes are located when the VT1.5, VT2, VT3, and VT6 all use a different number of columns? Recall that all VTGs are the same size regardless of the VTx flavor that they carry and that each VTG only contains only one type of VTx. Recall also that the VTxs within a VTG are column interleaved, that is, the first column of each of the VTxs within a VTG comes first, then the respective second columns, and so on. On top of this, all seven of the VTGs within the STS SPE are also column interleaved. This means that the first column from each of the VTGs occur sequentially within the STS SPE (after the path overhead column). The first byte contains the pointer bytes V1-V4, which also include the information on the size of the VT. Thus, the size of the VTxs within all the VTGs are known as soon as the first 7 bytes (following the first path overhead byte) are received. From this information, the pointer bytes for the rest of the VTxs can be located within each of the VTGs.

Table 2-3. Payload Types Specified by the Signal Label Byte C2

Code (in hex)	Payload Type
0x00	Unequipped (i.e., no path originating equipment)
0x01	Equipped, nonspecific (this can always be used)
0x02	Floating VT mode
0x04	Asynchronous mapping for DS3
0x12	Asynchronous mapping for 139.264Mbps
0x13	Mapping for ATM
0x14	Mapping for DQDB
0x15	Mapping for FDDI
0x16	PPP mapping with x^43+1 scrambling, RFC2615 [Malis+99]
0xCF	PPP mapping without scrambling, RFC1619 [Simpson94]

Of the four VT path overhead bytes, V5, J2, Z6, and Z7, only V5, J2, and 3 bits of Z7 have been specified so far. The eight bits constituting the V5 byte are used to provide the following functions: error performance monitoring (2 bits), a VT path remote error indication (1 bit), a path signal label (3 bits), and a path remote defect indication (1 bit, plus 3 bits from Z7).

2.4.6. Transparency

The layered model shown in Figure 2-17 is accompanied by the notion of transparent transport. That is, a lower layer treats the signal of the layer immediately above it as payload, and transports it without modification. In addition, different optical network elements may operate in different layers. A network element is said to be Section-, Line-, Tandem-, Connection-, or Path-terminating if it receives and processes the signal and the overhead at that layer, all layers below, and none above it (Figure 2-19). For example, the term Section Terminating Equipment (STE) denotes a network element that terminates and processes the physical and section layers of a SONET signal, but does not modify the line or path layers in any way. We have similar definitions for Line Terminating Equipment (LTE), Tandem Connection Terminating Equipment (TCTE), and Path Terminating Equipment (PTE). A PTE could be an STS PTE, a VT PTE or both, depending on the signals terminated.

Let us consider the following optical networking equipment and determine the SONET layer in which they operate: (a) a WDM system with OC-192 short reach interfaces, (b) an IP router with OC-48 or any OC-N interface, (c) an STS-level digital cross connect (called OXC, see Chapter 1), and (d) a T1 multiplexer with an OC-3 up link.

A WDM system with a SONET short reach interface on it will at least be terminating and regenerating the signal at the physical layer. It could provide section overhead functionality, if desired. However, since WDM systems came into existence after section level regenerators and there was a desire not to interfere with section layer communication and monitoring, most WDM systems do not perform STE functions.

An IP router with a SONET OC-N interface must be a PTE, since one of its functions is to pack IP packets into SONET STS paths. Note that many routers use a concatenated path signal of the same size as the SONET line, for example, an STS-48c path signal carried by an OC-48 optical signal. One might be tempted to say that in this case that the router uses the whole optical wavelength and hence requires a “lambda” (an entire wavelength) from the optical network. But this is not true! The router only requires that the path layer signal be transported transparently. The service it receives from the optical network does not change if the STS-48c path is later multiplexed with other STS-48c paths onto an OC-192 for efficient transport across the optical network.

An STS level OXC is a SONET path switch that can switch STS-1 and STS-Nc path signals but not VT level signals. In addition, such a switch does not alter the STS path signals in anyway. Note that this does not mean a switch of this type is not allowed to monitor path overhead, if it has the capability to do so. The act of switching STS paths from one SONET line to another requires de-multiplexing and re-multiplexing of STS paths onto SONET lines. Hence, the switch must terminate the SONET line overhead. In particular, it will have to rewrite the pointer bytes on egress. This cross-connect is a thus a SONET LTE.

A T1 multiplexer, by definition, has to take T1 (DS-1) signals and map them into VT1.5 paths, which then get packed into a VTG. The VTG is in turn packed into an STS-1 SPE, which is then carried in an OC-N. Hence, this equipment is a VT PTE. Note that it is also an STS PTE.

The SONET and SDH overhead bytes support different types of fault, performance and maintenance functionality. This does not mean, however, that all of this functionality must be implemented, present, and/or used in every deployment. Guidance on this comes from the standards with the notions of drop-side and line-side interfaces. A drop-side interface is an interface used for an intraoffice link. A line-side interface is one that is used for an interoffice link (i.e., between network elements). As an extreme example, consider the line order-wire capability (byte E2) for voice telephony communications between two LTEs. This capability is not meant for use over a drop side interface, for instance, to let two people within the same office communicate. The ANSI SONET standard [ANSI95a], in some cases, seems to rule out too much functionality on drop-side interfaces. For example, the section trace (J0), section DCC (D1-D3), and line DCC (D4-D12) are considered not applicable on the drop side. All these are methods provided by SONET for in-fiber communications between neighboring NEs, and they are essential for implementing neighbor discovery and port connectivity verification (see Chapter 6). Prohibiting these on the drop side may not have mattered when relatively few fibers were found in a central office. But with hundreds or thousands of fibers expected to terminate on new generation equipment, these capabilities become very important even on the drop side.

We continue our discussion of SONET and SDH in the next chapter, where we describe some of the more advanced features and recent additions to SONET/SDH for handling data traffic efficiently.