The Evolving Needs of Dispersion Testing in Today’s Next-Generation Networks

For many years, chromatic dispersion (CD) or polarization mode dispersion (PMD) testing was required for ultra-long-haul (ULH) submarine and terrestrial environments. While the network topology—comprised of the metro-edge network and the core or long-haul network—remains the same, a new topography for a broader CD and PMD testing landscape has emerged due to high-speed 40G and ROADM technology: the short-haul or metro and edge network. Known for its shorter distances, the absence of amplifiers and its once-low data rate, this type of network does not require CD and PMD testing. With new ROADM deployment and 10G/40G/100G, there is a shift in the requirements to test CD and PMD in metro areas. Today, CD and PMD testing need to be optimized for both metro and core networks.

The high-speed 10G/40G/100G market has also moved on in several ways in the last year. Coherent detection is making its way slowly and more robust modulation schemes such as differential phase-shift keying (DPSK) and differential quadrature phase-shift keying (DQPSK) are now common. Yet this explosion of new technology comes with its challenges. The most important one being profitability for telcos. They want to offer higher-speed services and still make money out of these services. How does this impact the testing landscape when it comes to fiber characterization?

1. Routes to Higher Transmission Speeds and Sufficient Accuracy for 10G, 40G and 100G Testing

Extreme accuracy is no longer needed in this new market reality. Higher accuracy for the very high acceptance parameters of earlier systems is not the main driver anymore. The first versions of 40 Gbit/s that came out to the market a few years ago were more or less faster versions of the 10 Gbit/s. Tolerances on optical impairment such as CD and PMD are becoming much more stringent, to the point where this would potentially jeopardize these higher-speed transmissions. The importance of very accurate testing became mandatory, and we have seen a number of units in the dispersion test world making more and more accurate measurements.

This physical limitation is one of the main reasons why it became mandatory to transmit data differently. First came duo-binary systems, which have evolved into DPSK and now into DQPSK. Higher tolerance to CD and PMD is one of the many advantages these transmission schemes. While it is still more susceptible than the lower-speed 10 Gbit/s, they are not as sensitive to it as the original 40 Gbit/s systems. Today, overtly high accuracy required for these earlier systems is no longer necessary. What is needed is rather sufficient accuracy for complete qualification of 10G, 40G and 100G networks while delivering better OPEX performance as well as faster and simpler testing.

We are also starting to see more advanced coherent detection-based systems, which promise to almost completely mitigate dispersion testing. However, while it reduces tolerance, it does not reduce the need for testing. Although we may get there one day, the first generations of these systems still suffer from dispersion and may get de-rated if dispersion is present. Nowadays, the most advanced system still requires dispersion especially since the main challenge remains: more dispersion means that high speed is not possible.

Although some measurement method offer higher accuracy than others, all test tools on the market have sufficient accuracy for 10G legacy, 40G DPSK/DQPSK as well as 40G/100G coherent detection. Therefore, test gear selection should not be based on this, but rather on OPEX reduction and ease of use as mentioned above. In the following section, we will highlight how a single-ended approach meets this challenge and how much OPEX can actually be saved.

2. Network Topology

Going back a few years, most systems were ring-based, with amplifiers and repeaters at strategic locations. The advent of reconfigurable optical add-drop multiplexers (ROADM) has created mesh networks, with amplification and compensation at every node and with data taking any of many possible routes from point A to point B.

As mentioned earlier, metro networks are typically shorter distances. Features in test equipment that were useful then, such as the capacity to characterize an entire ring (EDFA included) all at once, have little value in today’s environment. To create a dispersion map, one needs the granularity of the dispersion information, which represents the CD and PMD of every single section of the mesh individually.

This translates into testing of a lot of fibers in metro environments—that is the exact opposite of the goal to reduce truck rolls and OPEX. For example, this situation occurred in a mesh network like this one:

Having a single-ended instrument—a CD and PMD test tool that can characterize a section between two sites without having instruments at both ends—also means that many sections can be characterized in a few minutes instead of a few hours from a single location. As a result, an entire network can be characterized in 66% less time than any other traditional test methods. This greatly reduces truck rolls and OPEX, while increasing speed to deliver new services and reducing time-to-cash. Here is the truck roll scenario for the above network:

3. Business Model Evolution

Although end users want more bandwidth, they are not willing to pay a lot more per month for the service. Operators are looking for innovative ways of increasing the network bandwidth while reducing their OPEX. This takes many forms, such as outsourcing the installation and maintenance of the network as well as putting procedures into place to reduce truck rolls.

With such an increase in bandwidth demand, extra fiber deployments are required, which leads to more testing and more installation crews. This results in lowering the average expertise level of technicians while adding pressure on CAPEX and attempting to decrease the number of truck rolls.

This opens a door to test equipment vendors to bridge the knowledge gap and help entry-level technicians do the jobs faster and with higher success rates to decrease truck rolls and OPEX. Again, high accuracy is not required as much as before but getting it right the first time by avoiding complex setups and by putting intelligence in the equipment is where the value is in today’s test equipment.

Responding to the Industry’s Needs

With the market evolution, some test equipment vendors have met their customers’ needs. First, test sets must comply with ITU G.650.2 fiber testing standards and with EIA/TIA FOTP-175B for CD and EIA/TIA FOTP-243 for PMD. Vendors must also invest less in building overtly high-accuracy units that ultimately drive up instruments’ price with little true gain for the user. Vendors must develop test sets that respond to the industry’s increasing knowledge gap with entry-level field technicians by building intelligent and error-free setup units. Test sets should test multiple directions from one node, provide a one button test and one fiber connection, deliver one test file and require only one technician to do it all. This kind of efficiency results in cost savings.

Conclusion

There will always be a need for specialized dispersion instruments, which provides high accuracy and performs EDFA testing and distributed PMD measurements for bad section localization. However, since the bulk of the need follows market trends, users will seek test sets that deliver good enough accuracy, comply with industry standards like FOTP-243 and FOTP-175B, are easy to use, and allow for reduced OPEX and truck rolls.