1.2: SONET/SDH

SONET (Synchronous optical network) and SDH (synchronous digital hierarchy) is the foundation of modern telecommunications. By examining Figure 1.1.1, an analogy can be made with a river. The river is created by the confluence of many tributaries. In communications, each sender wants to communicate with one or more receivers. It is very inefficient to create a few billion independent data streams linking different senders and receivers. Instead, the data from different senders is multiplexed (combined) several times to create streams with a very high data rate. It is then also necessary to remove the right data from that stream (demultiplexing) so that the right data is sent to the right receiver. SONET/SDH enables different data streams with different base transmission rates to be efficiently multiplexed and demultiplexed.

At the same time SONET/SDH maintains a kind of "physical" compatibility with older technology (see Figure \(\PageIndex{1}\)). The original telephone network operated at a frequency of 8 kHz (period = 1 125 µs). As digital data became more significant, it became necessary to increase the data rate, so clock frequencies that are multiples of 8 kHz were introduced, allowing more data to be packed into the 125 µs window. For instance, the E1 standard carries 30 telephone channels (each 125 µs analog pulse is converted to a single 8 bit number) and two overhead channels, called an E1 frame. A single frame takes 125 µs to transmit, meaning that the base data rate is 2.048 Mb/s. E1 channels can be combined to create higher-order E numbers (E2, for instance has 4 E1 tributaries), which also have a frame that is 125 µs long. Likewise, SONET/SDH data frames are 125 µs long, but operate at much higher data rates (STM-1, the smallest SDH frame, is 155.520 Mb/s). This universal quality makes it easier for new technology to accommodate legacy equipment.

Figure \(\PageIndex{1}\): A visualisation of the relationship between data carrying capability and flexibility of different data carriers.

On the other hand, multiplexing creates an additional level of complexity. Imagine a phone call originating in New York and being received in Amsterdam: that tiny bit of data is multiplexed with many other phone calls and data streams so that it can be efficiently sent to Europe. However, at some point that one phone call needs to be demultiplexed. It might even be the only bit of data that needs to be removed from a much larger frame. Higher order E-frames do not support this. When E1 tributaries are multiplexed into E2, this is performed at the bit level, so a single byte has bits from each tributary. To demultiplex a single channel from a given E1 frame, all four E1 streams have to be separated, the relevant channel removed (from a single E1 frame), a new channel added, and then all four E1 tributaries are multiplexed again.

SONET/SDH accommodates efficient demultiplexing by including significantly more overhead. The payload of a SONET/SDH frame consists of virtual tributaries (denoted as VC), each of which are defined to match common lower data rates, as shown in Table \(\PageIndex{1}\).

Table \(\PageIndex{1}\): data rates for different standards and their corresponding virtual tributaries and their corresponding data rates.

Original standard and data rate	Virtual tributary and data rate
T1 (1.544 Mb/s)	VC-11 (1.728 Mb/s)
E1 (2.048 Mb/s)	VC-12 (2.304 Mb/s)
E3 (34.368 Mb/s)	VC-3 (48.960 Mb/s)
ATM (599.040 Mb/s)	VC-4-4C (601.344 Mb/s)

Notice that the virtual tributary data rate is always higher? This is for several reasons. First there is an internal pointer that keeps track of the location of the payload, even if it spreads across two frames. Second, the additional room allows for differences in clocking between different nodes by shifting bits within the frame.