New access services for end users increase demand for higher bandwidths in wide area networks. The step to transmission speeds of 40/43 Gbps corresponds to the next stage in the synchronous digital hierarchy (SDH).

The well-known technology takes on completely new perspectives and poses new challenges to measuring equipment. At last, telecommunications technology has entered a new round. After the boom and bust at the turn of the century, it seemed that the telecommunications industry was only focused on downsizing and consolidation. Only the metro and access sectors have shown slow but steady growth in the development of telecommunications technology.

DSL has been expanded continuously, with end users now having access to download speeds that are more than fifty times those of just a few years ago-and at prices that are affordable to private households. This development naturally has an effect on the planning of backbone network transmission capacity and technology. The next logical step is, therefore, being implemented the introduction of 40/43 Gbps technology.

To go from the 10 Gbps communications technology to the next higher level in SDH seems simple. However, it requires a lot of technological adjustments to achieve carrier-grade quality.

DWDM Wavelengths
Dense wavelength division multiplexing (DWDM) transmission systems for the wavelengths for the C and L bands specified by ITU-T provide up to 160 wavelengths as standard. Often, only the C band with just half of the wavelengths is used. Even this capacity is by no means fully utilized in many of today's networks.

From an economic point of view, it would seem better to make good use of the existing 10 Gb channel capacity for communications, rather than focusing on higher transmission speeds. However, the manufacturers of routers are the driving force for higher transmission rates. Cisco, for example, offers a switching capacity of 1.2 Tbps in a single-shelf configuration equipped with 16 line cards, each having a 40 Gbps interface. These router configurations can be cascaded almost without limits, which can potentially put a massive strain on the communications equipment.

There are still more challenges for communications equipment. SDH/SONET technology was established more than fifteen years ago, at a time when little or no thought had been given to the use of photonic networks, where optical communications channels could be transparently switched using optical cross connects.

Telecommunications is moving quickly toward this type of network, and the DWDM technology is aiding momentum. The signaling mechanisms used in SDH for a network operations center (NOC) are still based on end-to-end connections, and optical communication channels are not accounted for.

ITU has pressed forward with the standardization of the Optical Transport Network (OTN) for this reason, as described in Recommendation G.709. This OTN hierarchy has two main advantages over SDH: it defines both the optical communications channels and the associated signaling functions.

TCM is essential for routing the payload signals over the different parts of the network belonging to different providers. Path Monitoring (PM) allows for fault localization within individual network segments. The general communication channel (GCC) allows for signaling of the network elements involved with access to the ODU overhead.

And finally, a signaling function is also included for automatic protection switching (APS). Together, these functions allow monitoring and managing of optical communications signals across a number of network elements, eliminating the weaknesses of SDH/SONET.

FEC does lead to a distinct increase in the required transmission speed, for example, from 39.81312-43.01841 Mbps for an STM-256 signal. FEC uses most of this extra data, while the OTU3 framing information requires only a fraction of it. FEC is deliberately employed in order to realize longer transmission paths.

The loss budget in systems is the difference between the TX output power and the minimal RX input power, and it is used when planning the transmission path. The minimal receiver input power level is specified to achieve a transmission quality with a BER of 10-12. FEC allows acceptance of a higher insertion loss for the path, even if this means a BER of 10-7.

FEC corrects the errors and restores a BER of 10-12. The bit errors that this BER causes are also corrected by FEC. A 7% FEC results in an increased margin of some 5 dB, which is equivalent to a possible extension of the path by about 20 km.

The OTN hierarchy outlined by G.709 is not a new development. The technology did not achieve a breakthrough when it was originally defined some years ago. Globally, wide area networks were then mostly realized using SDH/SONET technology, and further development toward photonic networks had not taken place. Now, with the next stage in the transmission hierarchy achieved, OTN technology is in demand.

Source of impairments
The synchronous networks, described in the SDH/SONET standards and subsequently in the G.709 Recommendation for OTN, place extraordinarily high demands on the synchronization of the network elements and on the phase stability of the clock and data signals.

Synchronization in meshed networks is possible by feeding in a reference clock signal from a rubidium standard or using GPS on the network elements. Nevertheless, uncertainties in the clock signals still occur, particularly within the transmitting equipment, due to various physical mechanisms.

Jitter and wander are used to describe the usually periodic variations along the time axis of an ideal digital signal that can occur; for example, if the clock signal is unstable during sampling in the receiver. If this periodicity is below 10 Hz, it is called wander, as described in ITU-T Recommendation G.810. Jitter and wander have different causes.

Noise and crosstalk cause non-systematic jitter, mainly in high-frequency components. Pattern dependent, systematic jitter is generated by inter-symbol interference, which is the influence of a digital impulse on the impulses preceding and succeeding it.

Frame pattern jitter is the phenomenon of jitter that is generated by the signal structure caused by the non-scrambled part of the SDH or OTN signal (intrinsic jitter). With the increased number of overhead bytes at higher data rates, this effect becomes much more important in STM-256 OR OTU3 signals.

This type of jitter can be generated in the transmitter and also in the clock recovery stage of the receiver. It is characterized by a repetition rate of 8 Khz and its correlation to the SDH signal overhead.

Appropriate adaptation is required during multiplexing (mapping) of a constant data stream such as an SDH signal into an OTN frame. This is achieved by adding a byte that is not part of the payload signal (stuffing) or by shifting a payload signal byte into the next frame.

These bytes are removed during demultiplexing, the gaps closed, and the original clock frequency of the payload signal is reconstructed. Despite all these measures, low frequency jitter, or OTN mapping jitter, remains, which often has high amplitude.

Posing a Threat
The clock signal used to sample the transmitted signal at the ideal point, which is the center of each bit, is derived from the transmission signal itself in the clock recovery stage of the receiver. With low-frequency jitter on the signal, the derived clock signal follows the frequency variations and sampling is generally error free, as long as the jitter amplitudes are not too great.

With high-frequency jitter, the clock recovery circuits cannot follow the variations to the extent that signal sampling is free of errors. By the time the phase shift becomes larger than half the clock period (=0.5 UI, or unit interval) incorrect sampling of the payload signal is inevitable, resulting in bit errors.

Other factors may reduce the latitude in the decision threshold even further. It is possible that the payload signal cannot be synchronized (loss of frame, LOF) if the jitter amplitude is too high.

The performance of this oscillator must be optimal in order to measure the specified values with any degree of certainty, requiring costly development. According to ITU-T Recommendation O.172, the measuring equipment must demonstrate a maximum intrinsic jitter of 50 mUI in the relevant jitter frequency range.

The maximum level for high-frequency jitter for OTU3 is 100 mUI (10% of a unit interval).

Since peak-to-peak values are generally instantaneous, the rootmean square (RMS) value is calculated by integration. This in itself places high demand on the jitter analyzer, since its clock oscillator is also subject to thermal noise and phase impurities.

The performance of this oscillator must be optimal in order to measure the specified values with any degree of certainty, requiring costly development. According to ITU-T Recommendation O.172, the measuring equipment must demonstrate a maximum intrinsic jitter of 50 mUI in the relevant jitter frequency range.

The maximum tolerable jitter (MTJ) for an input stage is determined using a device that generates an OTN or SDH signal. Sinusoidal jitter is then applied to this payload signal at different frequencies. The measurement routine increases the jitter amplitude at each frequency until the bit errors, which occur, begin to exceed the specified limit value. This reduction in jitter is particularly necessary in meshed networks to prevent an accumulation of jitter over several segments of the network.

Measurement Accuracy
Fixed limits are specified for the intrinsic jitter of a jitter generator in ITU-T Recommendation O.172. Specifying the measurement accuracy of such jitter measuring equipment will ensure that the results obtained using different test equipment from different manufacturers is comparable.

The only way to determine the confidence interval of the results is to calibrate the receiver against a jitter-free generator. Producing such a device can be challenging. A method for determining the confidence interval and measurement accuracy of jitter measuring equipment is described in ITU-T Recommendation O.172, Appendix VII.

Taking the step to the next higher level in the communications hierarchy initially seems to be “just” to quadruple the bit rate. The 40/43 Gbps transmission speed demands much from transmission over optical fibers. Work is under way on new, multi-stage modulation methods in order to increase efficiency.

At the same time, new modules such as reconfigurable optical add/drop multiplexers (ROADMs) are making their way into network concepts. So, optical switching has once again become a hot topic. In new network concepts such as digital communications, analog parameters such as jitter are gaining importance, which also affect lower transmission speeds, because of the rapid development in the miniaturization of interface modules for economic reasons.

The high technical requirements still must be realized, thus measurement equipment must be adapted to the new technical environment to meet the new requirements.

Peter Winterling
The author is a senior solutions specialist, Optical Transport JDSU, Acterna Test & Measurement
(Germany)
vadmail@cybermedia.co.in