Optical communications: how to verify the real limits of data transmission with light

Data transmission using light encompasses very different technologies, from standard single-mode fibre to free-space systems. Before comparing figures or assessing a performance claim, it is therefore necessary to specify the medium through which the signal propagates and how it is detected. The medium and detection method determine which metrics are relevant and under what conditions a valid comparison can be made.

Single-mode optical fibre remains the backbone of global infrastructure. Its geometric and transmission characteristics are specifically defined, allowing equipment from different manufacturers to work together. This common basis matters when evaluating an improvement: an isolated speed figure does not, by itself, show whether a system can be integrated with the fibre and equipment used in a network. It is also necessary to know what kind of link was measured and under what configuration.

The Institute of Photonics and Quantum Electronics (IPQ) at Karlsruhe Institute of Technology (KIT) describes an evolution in the field: from a focus on long-haul transmission systems to data-centre interconnects. According to the institute, more than 50% of global data traffic is currently attributed to these short-reach links. This shift affects design priorities: raw capacity remains relevant, but energy efficiency and integration density are equally important when large numbers of devices must be connected in limited space.

Not all gigabits per second describe the same performance. Capacity per wavelength is a central metric, but it must be considered alongside transmission distance and bit error rate (BER). A figure without those details omits essential conditions: transmitting 100 Gbit/s over 10 metres in a laboratory is not equivalent to sustaining the same speed over 1,000 km with optical amplification. Reach and signal quality are part of how capacity should be interpreted, not minor details.

Chromatic dispersion and fibre nonlinearity impose physical limits that can be concealed by promotional figures. For a comparison to be useful, a report must state how transmission affects the signal and what target bit error rate is used. It is also worth checking that the quoted speeds refer to the same unit of comparison—for example, one wavelength—and not to the aggregate capacities of different systems. Without this context, two values expressed in Gbit/s may appear equivalent even though they come from incompatible tests.

Standardisation provides reference parameters. ITU-T Recommendation G.652, updated in 2024, specifies the characteristics of single-mode fibres whose zero-dispersion wavelength is around 1310 nm. These fibres can also be used in the 1550 nm region. The recommendation helps equipment from different suppliers interoperate and provides a common basis for describing the fibre. When assessing an improvement, it is therefore possible to ask whether it was measured under conditions compatible with this reference and what specific difference is proposed relative to the established baseline.

The International Telecommunication Union (ITU) publishes recommendations that define requirements for the safe and efficient operation of networks. They are not merely references for describing components: they also help specify the functions systems must perform in operating conditions and the aspects that should be monitored. When reviewing a performance claim, standards therefore provide context for distinguishing a laboratory figure from the behaviour that must be managed in a network.

ITU-T Recommendation G.697 defines optical monitoring for dense wavelength-division multiplexing (DWDM) systems. This monitoring is useful for configuration management, fault detection and tracking degradation. In an operational network, recognising changes or problems before they cause an outage is part of the practical evaluation of a system. However, these management functions do not always appear in academic papers, which may focus on capacity or a particular experimental result.

ITU-T Supplement G Suppl. 88 (10/2025) addresses requirements and transmission technologies for point-to-multipoint passive optical access systems operating above 50 Gbit/s per wavelength. It is a reference for understanding the direction of last-mile infrastructure and considers the difference between coherent and direct-detection technologies. Consulting the document helps place proposals within the framework of requirements relevant to this type of access; it does not, however, replace verification of the conditions and results for each system or component.

Published academic research, for example in Advances in Optics and Photonics, points to fundamental performance limits associated with fibre nonlinearities. These are not simply obstacles that can be removed by installing better hardware: they are consequences of the physics of light propagation. A capacity proposal must therefore be interpreted in light of the theoretical models that describe these limits and the conditions under which the result was obtained.

When assessing a new capacity record, the question is not only whether a test has exceeded a previous figure. It is also necessary to determine whether the comparison used equivalent parameters and how the result relates to theoretical limits. This check helps distinguish an incremental improvement from a possible paradigm shift, without confusing an experimental record with proof that physical limits have disappeared.

Safety is another dimension that should be considered separately from capacity. ITU-T Recommendation G.664 provides guidelines for optical safety at network interfaces, particularly for systems using high-power Raman amplification. Taking this reference into account broadens the assessment beyond speed: a claim about link performance does not, by itself, describe safe operating conditions. The recommendation provides context on this issue, while a specific analysis must take account of the technology and configuration described in the documentation.

Verification starts with primary documentation: the peer-reviewed paper or the complete technical report. Press releases and conference presentations may summarise a result without describing all test conditions, so they are not sufficient for rigorous comparison. The aim is to find enough information to reconstruct what was measured, how it was measured and how far the conclusion presented can be taken.

A useful checklist includes the length of the test fibre, the modulation type, the target bit error rate and energy consumption per bit. It should also clarify whether the system uses complex digital signal processing (DSP). Each item shapes the interpretation: distance describes the tested reach; modulation and DSP help characterise the configuration; and the error rate and energy consumption make it possible to examine quality and efficiency under the stated conditions. If these elements are missing, the comparison is limited.

Distinguishing levels of evidence

It is essential to distinguish between laboratory results, field demonstrations and operational deployments. A laboratory result may be obtained under ideal conditions that cannot be reproduced in a real network; it demonstrates what was possible in that configuration, not necessarily what is viable at scale. A field demonstration can validate integration in a real environment, although it does not by itself prove that the solution can be scaled up. An operational deployment provides the strongest evidence when it includes availability and maintenance metrics, because it allows sustained operation to be assessed rather than just a one-off test.

Reviewing sources for this article reveals a gap between standards documentation and specific research results. ITU-T standards are clearly defined, while articles in IEEE Xplore and Optica provide theoretical and design contexts but do not always present direct comparative data on current commercial products. A recommendation can therefore help establish criteria and describe requirements, but it cannot necessarily determine, on its own, which product offers better performance.

To verify a specific advance, readers need access to raw experimental data and the methods used. This information is not always available on an open-access basis. When the data or a sufficiently detailed description is missing, it is not possible to reconstruct the comparison fully; in that case, conclusions should be limited to what the source documents rather than extrapolated to other distances, equipment or conditions.

The rapid evolution of data-centre interconnects adds another difficulty: metrics from two years ago may already be obsolete. In addition, there is no universally accepted benchmark for energy efficiency in these specific links, which complicates direct comparison between research proposals. In this context, transparency in publishing experimental methods is decisive in helping others interpret and compare results. Available evidence makes it easier to assess a claim when conditions are stated, although it does not always resolve every difference between systems.