Introduction to SONET/ SDH SONET was developed in the United States through ANSI T1X1.5 committee.

ANSI work commenced in 1985 with the CCITT (now ITU) initiating a standardization effort in 1986. The US wanted a data rate close to 50Mbps. But the Europeans wanted the data rate to be around 150 Mbps. A compromise was reached and the US data rates were made subset of ITU specification, known formally as Synchronous Digital Hierarchy (SDH). SONET/SDH networks are configured as linear networks, where SONET/SDH nodes knows as Add Drop Multiplexers (ADMs) are hooked together in a line as shown in figure-1. There may be two or four fibers between the two consecutive ADMs with one set serving as protection or back up. Add/drop multiplexers (ADMs) are places where traffic enters and leaves. The traffic can be at various levels in the SONET/ SDH hierarchy (see Table-1). We will learn more about ADMs later.

Figure-1 Also SONET network elements can receive signals from a variety of facilities such as DS1, DS3, ATM, Internet, and LAN/MAN/WAN. They can also receive signals from a variety of network topologies. We will study how all this is done in subsequent sections. In addition SDH signals my also be connected with a SONET and vice versa. In this case, circuitry translates specific SDH information into its SONET equivalent, and vice versa. SONET/SDH Rates: The SONET frame in its electrical nature is called Synchronous Transport Signal-level N (STS-N). The SDH equivalent is called Synchronous Transport Module level N (STM-N). After conversion into optical pulses it is known as Optical Carrier level N. The line rates for different levels of SONET and SDH signals are shown in Table-1 below. Table-1:

Signal Designation Line Rate SONET STS-1 STS-3 STS-12 STS-48 STS-192 SDH STM-0 STM-1 STM-4 STM-16 STM-64 Optical OC-1 OC-3 OC-12 OC-18 OC-192 (Mbps)

51.85 155.52 622.08 2488.32 9953.28

You need not worry about the different levels of SONET /SDH at this stage. I had given detailed explanation of these levels later. I feel, to understand SDH easily, it is better to have knowledge of SONET initially. This is the reason I devoted major portion of this article to SONET. Except in terms of terminology there are no major differences between the two. But wherever there are differences I had pointed them out.

Basic SONET Frame Structure: The basic SONET frame is as shown in figure 2. This signal is known as Synchronous Transport Signal Level-1 (STS-1). It consists of 9 rows of 90 bytes i.e. a total of 810 bytes. It is transmitted from left to right and top to bottom. The two dimensional figure is just for convenience. Actual transmission takes place serially i.e. the left most byte in the top row is transmitted, then the second byte in the first row and so on. After the 90 byte in the first row the left most byte in the second row is transmitted and it goes on. One more point to be noted is that msb is transmitted first and the numbering of bits in a byte is as shown in figure-3. The frame length is 125s (i.e. 8000 frames per second). The STS-1 has a bit rate of 51.84Mbps. The frame for the lowest SDH rate STM-1 contains 270 columns by 9 rows. We will learn more about it later.
th

Figure-2

Figure-3

The first 3 columns of SONET frame are called Transport Overhead (TOH). The remaining 87 columns are calledSynchronous Payload Envelope (SPE). The first column of SPE is called Payload Overhead (POH). A point to be noted here is that every SONET frame repeats

every 125s no matter how fast the line speed gets. As line rate goes up SONET frame gets bigger, just sufficient to keep the frame rate at 8000 frames per second. SONET and SDH Interleaving: If three STS-1s are byte interleaved, the resulting frame is called STS-3. Byte interleaving means 1st byte of 1 STS-1 (called A1) is transmitted, then A1 byte of 2
rd st nd

STS-1, then A1 byte

of 3 STS-1 and so on. Now the resultant rate of the frame would be 3 times more. This multiplexing is carried out for all levels of SONET and SDH. Because of this, SONET/SDH maintains a frame rate of 125s. See figure-4 in which different colors for bytes of each frame are used to have clear understanding of multiplexing and how higher order frames are generated. Note that not all the overhead bytes in STS-1 frame are used in higher order frames. These are represented as X in the figure-4.

Figure-4 Transport Overhead: Let us go through each and every byte of Transport Overhead (TOH). I made sure that by the end of this section you will have fairly good idea of overheads in a SONET frame. Transport Overhead is divided into two parts -Section Overhead and Line Overhead. The first three rows inside the TOH constitute Section Overhead and the remaining six rows of TOH constitute Line Overhead (See figure-2).

Section Overhead:

Framing Bytes (A1 and A2): These bytes are used to indicate the start of SONET/SDH frame. A1 byte is 1111 0110 and A2 byte is 0010 1000. These values remain the same in all STS-1s in an STS-N. SDH uses the same values for framing. Section Trace (J0)/Section Growth: This byte formerly defined as STS-1 ID (C1), is redefined as Section Trace byte. It allows two connected sections to verify the connections between them by transmitting a sixteen-byte message. This message is transmitted in sixteen consecutive frames with first byte carried in first frame, second byte in second frame and so on. If no such section trace message is defined or being transmitted, then in STS-48 or lower bit rate the, J0 and each Z0 shall be set corresponding to its order of appearance in the STSN frame (i.e. J0 shall be set to 000000001, first Z0 to 0000010, second Z0 to 00000011 etc.) Where as in STS-192 frame each Z0 byte is set to the fixed pattern 11001100. Section BIP-8 (B1): B1 byte indicates bit error rate to the receiving terminal. This byte is known as Bit Interleaved Parity (BIP-8). The first bit in all the bytes in the previous frame are taken and then B1 is set so that the parity is even. Similarly all the other bits in B1 are set. The parity is calculated after scrambling and placed before scrambling. Scrambling is explained in later sections. The parity represented by this octet is the parity of the previous frame. It is used to estimate the bit error rate (BER) on the line. Note that the B1 parity is computed over all the bytes in the frame, no matter how large the frame. Because of this, the B1 byte does not provide a good BER estimation for large frames (perhaps STS-48 and larger) under adverse error conditions. SDH uses this byte for the same purpose. Orderwire (E1): The E1 byte is located in the first STS-1 of an STS-N. It is called Local Orderwire (LOW). The corresponding byte locations in the second through Nth STS-1s are currently undefined. This byte is used for a voice channel between two technicians as they installed and tested an optical link. It has a bit rate of 64kb/s. SDH uses this octet for the same purpose. Section User Channel (F1): The F1 byte is located in the first STS-1 of an STS-N, and is used by the network provider. The corresponding byte locations in the second through Nth STS-1s are currently undefined. This byte is passed from Section to Section within a Line and can be read, written, or both at each Section Terminating Equipment (STE) in that line. The use of this function is optional. SDH also uses this byte for the same purpose. Section Data Communication Channel (D1, D2 and D3): These are the bytes, which form communication channel. These bytes are defined only for first STS-1 of an STS-N frame. These three bytes are considered as one 192-kb/s, message-based channel for alarms, maintenance, control, monitoring, administering and other communication needs between

STE. This channel is used for internally generated, externally generated and supplier-specific messages. SDH uses this channel for the same purpose. Line Overhead: Pointers (H1 and H2): The processing of H1 and H2 bytes in SONET and SDH is a beautiful concept. Consider figure-2 which describes the STS-1 frame. The SPE can be floating in a SONET frame. It can start in one frame and end in the next frame. Now these two bytes are allocated to a pointer that indicates the offset in bytes between the pointer and the first byte of the STS SPE. The pointer bytes are used in all STS-1s within an STS-N to align the STS-1 Transport Overheads in the STS-N, and to perform frequency justification. You need not worry about this concept as of now as I am going to explain it much more detail in later sections. SDH handles these pointer bytes in the same way. Pointer Action Byte (H3): This byte is used when negative frequency justification is carried out. The value carried by H3 is not defined when there is no negative frequency justification. SDH handles this byte in the same way. Line BIP-8 (B2): The operation of this B2 byte is same as that of B1 byte in the SOH except that B2 is calculated over Line Overhead and Synchronous Payload Envelope of the previous frame before scrambling and placed in the current STS-1 frame before scrambling. SDH uses this byte for the same purpose. Automatic Protection Switching (APS) Channel (K1, K2): In the beginning of this article, I told that a set of fibers is used for protection. These K1 and K2 are the bytes, which are transmitted over these protection channels for Automatic Protection Switching (APS) signaling between line level entities. These bytes are defined only for first STS-1 of an STS-N. In the remaining STS-1s it is undefined. These bytes are used to indicate a number of defects, alarms etc. detected at the receiving terminal back to the corresponding transmitting terminal through protection channels. SDH uses these bytes for the same purpose. There is lot more explanation to be done on this concept of APS. Ill discuss this in later sections. Line Data Communication Channel (D4-D12): These bytes form a communication channel to send administrative messages just as D1 to D3. These nine bytes are considered as one 576-kb/s, message-based channel for alarms, maintenance, control, monitoring,

administering and other communication needs. This channel is available for internally generated, externally generated and supplier-specific messages. These bytes are defined only for STS-1 number 1 of an STS-N signal. SDH uses these bytes for the same purpose but with additional codes.

Synchronization Status (S1): This byte is allocated for transporting synchronization status messages. S1 is defined only for first STS-1 of an STS-N signal. Currently only bits 5-8 of S1 are used to transport synchronization status messages. Bits 1-4 are undefined. These messages contain clock quality labels that allow a SONET NE to select the most suitable synchronization reference from the set of available references. The purpose of these messages is to allow SONET NEs to reconfigure their synchronization references autonomously while avoiding the creation of timing loops. Let me give a few examples for bits 5-8 in S1. Bits 5-8 are 0001 for stratum 1 traceable, 0111 for stratum 2 traceable, 0000 Synchronized traceability unknown etc. SDH uses this byte for the same purpose Growth (Z1): Z1 byte is located in second through Nth STS-1s of an STS-N. This byte is undefined. STS-1 REI (M0): The M0 byte is defined only for the STS-1 in an OC-1 or STS-1 electrical signal. Bits 5 through 8 of the M0 byte are allocated for a Line Remote Error Indication function (REI-L), which conveys the error count detected by LTE (using the B2 code) back to its peer LTE. Bits 1 through 4 of the M0 byte are currently undefined. The error count shall be a binary number from zero (i.e., 0000) to 8 (i.e., 1000). The remaining seven values represented by the four REI-L bits (i.e., 1001 through 1111) shall not be transmitted, and shall be interpreted by receiving LTE as zero errors. Since there is no rate in SDH equivalent to STS-1, SDH does not define an M0 value for this byte. STS-N REI (M1): The M1 byte is located in the third STS-1 in an STS-N (N 3) and is used for Line Remote Error Indication (REI-L). The entire M1 byte is used to convey the count of errors detected by the Line BIP-8 (B2) byte. This count has (8 times N) + 1 legal values, namely 0 to 8N errors. For rates below STS-48, the remaining possible 255-(8 times N) values are interpreted as zero errors. For the STS-48 and STS-192 rates, if the line BIP-8 detects greater than 255 errors, the line REI will relay a count of 255 errors. SDH uses this byte for the same purpose. Growth (Z2): These bytes are allocated for future growth, and their use is currently undefined. Note that STS-1 signal does not contain a Z2 byte. Orderwire (E2): This byte has the same purpose for line entities as the E1 byte has for section entities. It is called Express Orderwire (EOW) channel. The corresponding bytes in the second through the Nth STS-1s of an STS-N frame are currently undefined. SDH uses this byte for the same purpose. Before getting into Synchronous Payload Envelope (SPE), I feel its better to know about scrambling.

Every add/drop multiplexer sample incoming bits according to a particular clock frequency. Now this clock frequency is recovered by using transitions between 1s and 0s in the incoming OC-N signal. Suppose, incoming bit stream contains long strings of all 1s or all 0s. Then clock recovery would be difficult. So to enable clock recovery at the receiver such long strings of all 1s or 0s are avoided. This is achieved by a process called Scrambling. Scrambler is designed as shown in the figure-5 below.

Figure-5 It is a frame synchronous scrambler of sequence length 127. The generating polynomial is 6 7 1+x +x . The scrambler shall be reset to 1111111 on the most significant byte following Z0 byte in the Nth STS-1. That bit and all subsequent bits to be scrambled shall be added, 7 modulo 2, to the output from the x position of the scrambler, as shown in Figure 5. The framing bytes A1 and A2, Section Trace byte J0 and Section Growth byte Z0 are not scrambled to avoid possibility that bytes in the frame might duplicate A1/A2 and cause an error in framing. The receiver searches for A1/A2 bits pattern in multiple consecutive frames, allowing the receiver to gain bit and byte synchronization. Once bit synchronization is gained, everything is done, from there on, on byte boundaries SONET/SDH is byte synchronous, not bit synchronous. An identical operation called descrambling is done at the receiver to retrieve the bits. I already said in one of the previous sections that the STS SPE would be floating and the pointers give the position from where it is going to start. The phase differences between Transport Overhead and SPE must be accommodated. Also SONET/SDH systems are synchronous systems, i.e., all the clocks are supposed to be same, but in reality they are not. There shall always be small differences. Even when all the clocks are same there can be jitter, which must also be accommodated. Data can come into a device slower or faster than it is transmitted out at the other side. So something has to be done to adjust the differences between the transmit and receive clocks. This is where pointer action bytes H1, H2, and H3 come in.

H1 and H2 bytes are called pointer bytes. Consider H1 and H2 as shown in figure-6. Bits 1 to 4 are called New Data Flag (NDF) bits represented by Ns. It is set to 0110 for normal operation. Bits 5 and 6 are currently undefined in SONET but are used in SDH, which we will study when we go through SDH. Bits 7 to 16 carry the pointer value.

Figure-6 The value of the bits 7 to 16 can vary from 0 to 782. A value of 0 indicates that SPE starts at the first byte immediately after H3 byte. If the pointer value is 1 the payload starts at the second byte after H3 etc. Figure 6 shows the layout of H1 and H2 pointer. For the time being ignore the meaning of I and D labels. Figure 7 below shows locations of SPE for different values of pointer. 1 2 3 4 5 6 89 90

H1

H2

H3

0 87

1 88

85

86

521 522 523

781 H1 H2 H3 0 1

782

Negative stuff Opportunity Positive stuff opportunity Figure 7

Now let us see why and how H3 is used. Suppose the incoming clock is faster than the outgoing clock. Then an extra byte is accumulated in our receive buffer, compared to what we can transmit. Now this extra byte is put into H3 location. So when we transmit one SONET frame of 810 bytes, we actually transmit 784 bytes of payload (86 columns times 9 rows, plus one H3 octet), rather than 783 bytes of payload. A similar problem occurs if the incoming clock is slower than the outgoing clock. Then there will be a deficit in the receiver buffer. To overcome this problem a stuff byte in the location after the H3 byte is sent. Moving of SPE backwards is called negative justification and moving it forward is called positive justification. In figure-6 pointer bits are labeled as IDIDIDIDID. I indicates increment and D indicates decrement. With the help of these I and D bits we can make positive or negative justification. Let us go through how these positive and negative justifications are done. Sonet equipment will have a register using which it compares the present pointer bits with the previous bits. If I bits are inverted then pointer value is incremented (positive justification) or if D bits are inverted then pointer value is decremented (negative justification). By changing the NDF value also new pointer value can be introduced. We shall discuss all this in more detail.

Positive Justification: Frame Status Normal Frame Inverted I bits New Ptr Value New Ptr Value Normal Frame Figure-8 New Data Flag 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 Unused Bits X X X X X X X X X X I 0 1 0 0 0 D 0 0 0 0 0 I 1 0 1 1 1 D 1 1 1 1 1 I 0 1 0 0 0 D 1 1 1 1 1 I 0 1 0 0 0 D 1 1 1 1 1 I 1 0 1 1 1 D 0 0 1 1 1

Figure-9 Consider Figures 8 and 9 for the explanation that follows. Let the pointer value be 214. Its decimal equivalent is 0011010110 which is shown in first row in figure-8. Now at the frame where positive justification should be done the I bits are inverted which is shown in the second row in figure-8. It is in this frame that the byte after the H3 is stuffed. As soon as the receiver detects that the I bits are inverted, it discards the byte after H3 as the content is meaningless-but it usually contains all 0s and also the pointer value is incremented at the receiver side to point to the adjusted payload. In the next frame the receiver receives the incremented pointer value. This new pointer value is repeated for the next frame as well as for the fourth frame. The new pointer value received in the fourth frame is available for pointer adjustment again where as that received in the previous two frames is not available for adjustment. This is because a new pointer value is accepted only when it is received for three consecutive frames. Negative Justification: Frame Status Normal Frame Inverted D bits New Ptr Value New Ptr Value Norm Frame Figure-10 New Data Flag 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 Unused Bits X X X X X X X X X X I 0 0 0 0 0 D 0 1 0 0 0 I 1 1 1 1 1 D 1 0 1 1 1 I 0 0 0 0 0 D 1 0 1 1 1 I 0 0 0 0 0 D 1 0 1 1 1 I 1 1 0 0 0 D 0 1 1 1 1

Figure-11

Consider Figures 10 and 11 for the explanation that follows. It is very much similar to positive justification. Let the pointer value be 214. Its decimal equivalent is shown in first row in figure8. Now at the frame where negative justification should be done the D bits are inverted which is as shown in the second row in figure-10. It is in this frame that the byte after the H3 is used to carry a data byte. As soon as the receiver detects that the D bits are inverted, it immediately takes up H3 as it contains some relevant data and also the pointer value is decremented at the received SONET equipment to point to the adjusted payload. In the next frame the receiver receives the decremented pointer value. This new pointer value is repeated for the next frame as well as for the fourth frame. The new pointer value received in the fourth frame is available for pointer adjustment again. Pointer Adjustment Using NDF: Frame Status Normal Frame NDF Set New Ptr Value New Ptr Value Normal Frame New Data Flag 0 1 1 0 1 0 0 1 0 1 1 0 0 1 1 0 0 1 1 0 Unused Bits X X X X X X X X X X I 0 0 0 0 0 D 0 0 0 0 0 I 0 0 0 0 0 D 1 1 1 1 1 I 0 0 0 0 0 D 1 1 1 1 1 I 0 0 0 0 0 D 1 1 1 1 1 I 0 1 1 1 1 D 1 0 0 0 0

Figure-12 Till now we have discussed incrementing or decrementing of pointer value only by 1. What if we want to adjust pointer value by more than 1. This can be achieved using NDF. To make adjustments using NDF, it's bits must be set to1001. The frame in which NDF is set new pointer value is introduced immediately. Here there is no need for inversion of I or D bits. Pointer adjustments using NDF cannot occur more often than every fourth frame in SONET. This is so because the new pointer value is accepted only if it is received for three consecutive frames. Another reason for the limit of every fourth SONET frame is because of two special cases. Special Cases: In the positive pointer adjustment case suppose the pointer is changed from 782 to zero. Here you can see that the pointer in the frame following the frame with the NDF inverted will point to the same SPE as the (implied) pointer of the frame with the inverted NDF. This means that the system has to ignore the pointer for one SONET frame and the first valid pointer will be the third SONET frame (the frame with the NDF inverted is one, the next frame is two, and the third frame has a valid pointer). A similar special case occurs when the pointer rolls over backwards from zero to 782. The zero value pointer is pointing at the byte after the H3 byte. The next frame has the NDF set and the H3 byte is used for data. The resultant pointer is pointing to the H3 byte. The beginning of the next SPE is the last byte in row 3 of the SONET frame after the frame with the NDF inverted. However, the pointer in that frame cannot point backwards in the frame, so the SONET equipment has to handle this situation as a special case. The pointer in that SONET frame points to the last byte in row 3 of the next (third) SONET frame. This can be confusing but if you draw pictures similar to Figure 9 and Figure 11 you shall see how it works. We will be processing pointers in a different way when we are using concatenated structures, which we will be discussing later.

Understanding SONET By K. Surya Prakash