SONET
In 1984, the U.S. telephone carriers (regional Bells) and international phone companies sought a standard to replace their proprietary fiber infrastructure. Deciding upon a standard would allow them all to buy equipment from the same vendor pool, bringing down costs as well as expanding the amount of available equipment.
This task fell to the Exchange Carriers Standards Association (ECSA—now ATIS, the Alliance for Telecommunications Industry Solutions), which developed SONET for connecting disparate fiber networks.
SONET is a circuit-switched method of multiplexing over a fiber network. The multiplexing scheme is called "byte-interleaved." This sends time-divisioned signals from multiple sources over the same medium. The base signal in SONET multiplexing is called transport signal level-1, or STS-1. It's a 51.84Mbps signal. Multiplexing takes place at multiples of STS-1 (see Table 2).
Table 2 SONET Multiplexing Frequencies
|
Name |
Frequency |
Comparative DDS Bandwidth |
|
STS-1, 51.840 Mbps |
28 DS1s |
(1 DS3 ) |
|
STS-3, 155.520 Mbps |
84 DS1s |
(3 DS3s) |
|
STS-12, 622.080 Mbps |
336 DS1s |
(12 DS3s) |
|
STS-48, 2488.320 Mbps |
1344 DS1s |
(48 DS3s) |
|
STS-192, 9953.280 Mbps |
5376 DS1s |
(192 DS3s) |
The STS frame is 90 bytes wide and 8 bytes deep. With 8 bits per byte sent 8,000 times a second, this is 51,840,000 bits per second, or 51.840Mbps.
Although the Plesiochronous Digital Hierarchy (PDH) is the traditional method of multiplexing channels, it suffers from a number of problems. First, as its name implies, it's not completely synchronous. At rates above the DS1, it's almost synchronous. There was no standardization with international telephone systems. And interoperability between different vendors' equipment was not guaranteed, even though they were built to the same standards.
Multiplexing different data streams together that do not have a single time reference can be a daunting task. This task is complicated by the fact that the data travels at different rates depending on temperature and distance. These factors and more mean that multiplexed DS1 data streams do not align. DS-2 and DS3 multiplexed data streams are therefore not synchronous.
Telco engineers worked around this asynchronous transmission problem using "bit stuffing." Adding additional bits here and there makes the multiplexed data streams difficult to address directly. To access a DS0 in a DS3, the DS3 first has to be demultiplexed down to four DS2s. Then the DS2 that contains the call that needs to be modified is demultiplexed down to seven DS1s. At the DS1 rate, individual DS0s can be addressed. This process is expensive, both in terms of processing necessary to drop or add calls and in terms of equipment necessary to do that processing.
Synchronous transmission is, by definition, synchronized and does not use bit stuffing. Therefore, it is not as complicated a task to strip one data stream from a multiplexed pipe.
In 1987, Bellcore proposed SONET as a replacement for PDH to the CCITT. This original design of SONET was rejected by the CCITT. One of the main problems was that SONET was originally defined to transmit at 49.9Mbps, which works great with North American DS3s, but it does not accommodate European CEPT standards.
See the digital hierarchy in Figure 6.
){kind=link}
Figure 6 Plesiochronous Digital Hierarchy and comparative bit rates.
You can compare this to the SONET/Synchronous Digital Hierarchy (SDH) in Figure 7.
){kind=link}
Figure 7 The SONET/synchronous digital hierarchy.
Believe it or not, the Americans, the Europeans, and the Japanese were each able to make concessions, and the CCITT (Comité Consultatif International Téléphonique et Télégraphique—now, after a 1992 reorganization, known as ITU, the International Telecommunication Union) produced a nearly unified Synchronous Digital Hierarchy (SDH) standard. Two major concessions were made to reach consensus. First, the Americans increased the basic rate of SONET from 49.92Mbps to 51.84Mbps. Second, the Europeans gave up on having their CEPT-2 and CEPT-3 PDH rates directly supported.
SDH is the proper international name for SONET/SDH networks. Although many people use SONET and SDH interchangeably, there are minor differences between the two, and SONET and SDH networks do not interoperate without conversion equipment. This equipment is trivial compared to that which was necessary for the PDH network.
SONET differs from the legacy PDH in many ways. Most important is that SONET synchronizes at bit rates faster than DS1. The capability to address individual DS0 channels without first having to demultiplex a DS3 several times is a multibillion-dollar savings for the telco. Individual calls could be dropped or added without all the additional processing and equipment at each central office.
SONET transmits using a highly stable reference-clocking source for the entire SONET network. Because there is no need to synchronize clocks or align data streams, there is no need to do bit stuffing. In instances in which data might vary, pointers—similar in concept to pointers in the C programming language—in the SONET frame header are used to determine the exact offset of the data within the frame payload. This allows, for example, data in a DS1 to be directly accessed without first demultiplexing.
SONET itself is nothing more than a fast direct method of transporting data that is costly but resistant to delays. SONET can carry many upper-level protocols: legacy PDH DS-3 and CEPT-4 circuits, FDDI, and, of course, ATM. The most important, at this point in time, is the support for legacy PDH technologies. This allows the telecom companies to deploy SONET while maintaining connectivity for their customer base, thus allowing a gradual phaseout of the aging PDH technology.