Turning Up 40G Submarine Networks—The Testing Challenges

From the planning stage to the deployment of an undersea fiber-optic cable, a considerable amount of time, money and resources are invested to ensure the success of a submarine network. Yet one point that is often overlooked during this process, and which can lead to unfortunate delays, is the final acceptance test of the fiber. Simply put, it comes down to making sure that the fiber can deliver on the bandwidth promise. This final acceptance test is often ignored since legacy data rates are considered as being extremely challenging to transmit. Yet with the deployment of 40G, newer and more serious challenges are at hand, and in order to face them, they must be understood and properly prepared for. To do so, some critical physical-layer tests must be performed, including dispersion testing, such as chromatic dispersion (CD) and polarization mode dispersion (PMD). Tests that were considered as being of mild importance at 10G have become critical at 40G transmissions. Once these parameters have been fully qualified and optimized, the system turn-up tests must then be performed; this includes signal power level and optical signal-to-noise ratio (OSNR) measurements, which are important in both repeated transoceanic or festoon links. Finally, the network end-to-end level qualification must be done and all the relevant SONET/SDH tests are performed, including tests such as bit error rate (BER).

This article reviews the challenges brought forth by the advent of 40G transmission and describes the importance of each test, as well as where and how they should be performed; what’s more, it examines a case study of a trans-Atlantic dispersion acceptance test that was performed in 2010.

Context

As the demand for bandwidth continues to grow at a tremendous pace in submarine networks, service providers have no choice but to adapt their networks so that they too can offer more services to their customers. Yet with new technologies rapidly approaching, service providers are faced with difficult choices, and in this competitive context, upgrading from a 10G to a 40G advanced modulation format becomes a necessity—but not at any price.

The three main drivers behind each upgrade are: (1) service providers want to use existing fibers and network elements to maximize revenues from infrastructure that is already in place; (2) they want to improve operation flexibility, which translates into more efficient provisioning and increased revenues from new services to their customers; and (3) they want to increase the density of the network elements to optimize central office space and power consumption. However, upgrading from 10G to 40G, or even to 100G, has a significant effect on the tolerance to optical impairments, such as the level of noise with the signal, as well as the risk of CD and PMD.

The effects of CD and PMD have been known for decades, yet testing these parameters on OC-192/STM-64 is not always done. But now, with the current penetration of 40G, CD and PMD are becoming an even bigger issue. New 40G transmission rates come in many flavors, yet even the most resistant OC-768/STM-256 (40G) can only tolerate a fraction of what OC-192/STM-64 (10G) can accept—both in CD and PMD. The typical acceptable CD rate is around 10% of the 10G tolerance levels, while PMD tolerance is typically at 25%. This means that PMD values that were previously considered as acceptable, may not be anymore.

Moreover, CD compensation schemes were effective at 10G, yet the residual compensation is much too high for 40G transmission rates, and more specifically, PMD was never compensated—either it was ‘hoped to be good’, or it was tested more often and when it was too high, the fiber was limited to lower speeds. But CD was compensated via simple dispersion-compensating modules (DCMs), which are deployed automatically and installed by system vendors. On the other hand, accurately testing for this parameter allows one to fine-tune each stage of compensation, as well as to perform (if required) end-to-end compensation adjustments to remove excess residual dispersion.

The following case study highlights the importance of CD and PMD measurements in 40G submarine deployments.

Case Study

This case study examines the submarine network from the UK to New York, which was tested for dispersion parameters. The one-way trip (UK-USA) was 6008 km in length and was comprised of 144 erbium-doped fiber amplifiers (EDFAs). Two fibers were tested: A4 and A5. The following tests were performed:

• CD and PMD on fiber A4, CD was performed twice to validate repeatability
• CD and PMD on fiber A5, PMD was performed twice to validate repeatability

The results were quite revealing:

These results were in the expected range: CD compensated properly and therefore, was barely low enough for 10G, but it was clearly out of specifications for any 40G. PMD was below the 10G tolerance, but one of them, fiber 4A, did not pass the 40G typical PMD tolerance (2.5 ps being the typical value).

To prove the accuracy of the test measurements, loopback tests were performed. CD being linear, looping it onto itself should result in doubling the value. PMD being stochastic, the total expected value would be the quadratic sum. The results are as follows:

Both loopback results are within the expected range of value (considering the test equipment uncertainty), confirming the validity of the single direction tests and therefore the conclusions. This clearly proves the importance of CD and PMD tests for 40G submarine deployments, regardless of whether it operates at 10G flawlessly. The second parameter to be monitored in 40G networks is OSNR, which is an excellent way to provide a representation of all the optical noise effects. Typically, there is a direct relationship between the OSNR and the BER; therefore, higher OSNR leads to lower BER.

OSNR is generally measured by using the interpolation method, as recommended in IEC 61280-2-9. This method is based on the assumption that noise level is generally flat between and under adjacent peaks. Unfortunately, the interpolation approach is not possible when the peaks are very closely spaced—as with large 40G signals or ultra-dense wavelength-division multiplexers (UDWDM)—and crosstalk becomes dominant over ASE noise, see figure 1.

In this case, an innovative method, based on advanced analysis using fundamental differences between signal and noise, namely polarization and spectral characteristics, allows an adequate measurement of the noise within the channel—without having to turn off the transmitter. This approach is referred to as in-band OSNR measurement and takes advantage of built-in polarization diversity detection combined with a polarization controller to measure the power versus the wavelength on two polarization axes for multiple polarization states; thereby discriminating between the polarized and non-polarized power and isolating the noise from the signal inside the channel. Relying on sophisticated algorithms that work concurrently on the signal shape enables the user to automatically achieve accurate and repeatable OSNR measurements in 40G networks. Finally, qualifying the OC-768/STM-256 40G payload is an essential part of turning up 40G submarine links and is typically conducted through the end-to-end, BER test (BERT) across the network. With BERT, service providers can ensure that all new 40G circuits in service are proven capable of handling any of the demanding services that may be placed on them in the future. Therefore, these test routines, when performed without the proper test equipment, can prove complex and cumbersome to perform.

Conclusion

With the deployment of 40G and up, new challenges are at hand, as discussed above. In order to meet those challenges, they must be well understood and planned for. Some critical physical-layer tests include dispersion testing, such as chromatic dispersion (CD) and polarization mode dispersion (PMD). Furthermore, network end-to-end level qualification is required where all the relevant SONET/SDH tests are performed, including tests such as bit error rate. These are only some of the tests that at 10G are important, yet become mission critical at 40G transmissions.