Page 78 - Optical Switching And Networking Handbook
P. 78

04_200023_CH03/Batesx  1/17/01 9:42 AM  Page 63










                    SONET                                                                         63


                                       Why Bother Synchronizing?


                                       In the past, transmission systems have been primarily asynchro-
                                       nous. Each terminal device in the network was timed independently.
                                       In a digital synchronous transmission system, clocking is all-
                                       important. Clocking uses a series of pulses to keep the bit rate con-
                                       stant and to help recover the 1s and 0s from the data stream.
                                       Because these past clocks were timed independently, large varia-
                                       tions occurred in the clock rate, making it extremely difficult (if not
                                       impossible) to extract and recover the data. A DS-1 stream operates
                                       at 1.544 Mbps   150 pulses per second (pps), whereas a DS-3 stream
                                       operates at 44.736 Mbps   1789 pps. This difference means that one
                                       DS-1 stream may be transmitting at up to 300 pps more or less than
                                       another (assuming that one DS-1 signal is at -150 pps and the other
                                       is at  150 pps). The differences can make it difficult to determine
                                       the actual data across a common receiver.
                                         Earlier we discussed the asynchronous method of multiplexing a
                                       DS-3 stream. We saw that bit stuffing was used. Therefore, the
                                       method of synchronously multiplexing in the SONET architecture
                                       provides for better efficiency and problem resolution. Using SONET,
                                       the average frequency of all the clocks will be the same. Every clock
                                       can be traced to a common reference, which is highly reliable and
                                       stable. Bit stuffing can be eliminated in the preparation of the STS-
                                       1 signal, and therefore, the lower speed signals are more readily
                                       accessible. The benefits outweigh the possible overhead associated
                                       with the SONET multiplexing scheme. In SONET, the hierarchy of
                                       clocking follows the master-slave clocking architecture. Higher-level
                                       (stratum 1) clocks will feed the timing across the network to lower-
                                       level devices. Any jitter, phase shifts, or drifting by the clocks can be
                                       accommodated through the use of pointers in the SONET overhead.
                                       The internal clock in a SONET multiplexer also may draw its timing
                                       from a building-integrated timing system (BITS) used by switching
                                       systems and other devices. This terminal will then serve as the mas-
                                       ter for other SONET nodes downstream, providing timing on its out-
                                       going signal. The receiving SONET equipment will act in a slave
                                       mode (loop timing) with their internal clocks timed by the incoming
                                       signal. The standard specifies that SONET equipment must be able
   73   74   75   76   77   78   79   80   81   82   83