The age of ‘Big Data’ and the power of real-time data analytics means organizations are increasingly turning to enterprise networks to handle the need for ever increasing capacity of low latency, secure and stable connectivity.1 An important part of any datacom system and achieving optimal performance in these areas is the use of spectral filters as part of the datacom network.

Importance of Optical Filters

Datacom, in particular enterprise networks and data centers where large volumes of data are being transmitted, need to be able to handle large bandwidths of data transfer while maintaining data integrity.

Now, most datacom applications make use of optical fibers for data communication due to the enhanced transmission speeds and bandwidth over traditional copper wiring. Optical filters help optical fiber communication achieve its full potential by allowing for multiplexing in digital center interconnects.2 Datacom optical filters can be used to suppress unwanted background and allow for selective transmission only of certain transmission channels, particularly where multiple signals are being sent over a single fiber and cross-channel communication may be a concern.

Multi-channel Transmission

There are a number of ways of implementing multi-channel transmission over single fibers within datacom environments. The motivation for doing this is to increase the amount of information that can be transmitted but also allow for redundancy in data transfer, to improve communication robustness, and flexibility in the types of information that can be transferred.

One of the most common implementations in datacom is wavelength division multiplexing (WDM).3 In WDM, multiple data streams are transmitted in different frequency/wavelength ranges through a single optical fiber.

Three of the main wavelength regions that are used in WDM are the O-band (1260-1360nm), C-band (1530-1565nm) and L-band (1565-1625nm).

For data center applications, it is often important that any filters – such as broadband filters that allow all of the O, C and L-band through together – are highly compact to not increase the footprint of devices. Broadband filters help to remove interference from any signals outside of the desired signal transmission ranges and the use of very thin filters also helps maintain high-speed data transmission as well.

Datacom vs Telecom Requirements

There are a few key differences between datacom and telecom WDM. Datacom needs much higher data transfer rates with lower latency than telecoms requirements so filters often need to be broader bandwidth, lower loss and thinner to ensure this.

Critical Specifications

Central wavelength and passband width: When choosing a bandpass filter to only let through a selected band, it is imperative the central wavelength and filter bandwidth or passband width are selected correctly. Typically central wavelengths for filters for a single band are as follows: O-band (1310 nm), C-band (1550 nm), and L-band (1590 nm). As the entire O-band region is usually 1260-1360nm, it is also necessary for the passband width of the filter to be at least 40-60 nm to avoid blocking transmission of signals in certain bandwidth regions.

Low loss: Low loss or high transmission filters are essential in high-speed data environments as poor filter transmission can end up degrading the overall signal amplitude and quality, ultimately resulting in the loss of signals. Insertion loss is defined as being losses from the introduction of devices, such as optical filters, into the transmission path. For example, TIA 568.3-D specifies a maximum insertion loss value for fiber connectors at 0.75 dB.

High Isolation: Isolation determines how much ‘crosstalk’ there is between channels in the densely packed data environments. Typical isolation levels required in these filters are greater than 30-40 dB to avoid contamination of signals between different channels.

Low Polarization Dependent Loss (PDL): Optical components often have different transmission properties for different polarizations, which is problematic for datacom where information is often encoded using different polarizations of the transmitted light. Ultimately, if the losses for different polarizations are very different, there can be a significant impact on signal integrity and so low PDLs across the full operation bandwidth of the filter are important.

Design Challenges and Technological Innovations

Many datacom filters need to be compliant with stringent specifications to ensure a high signal-to-noise ratio is maintained and avoid information loss. Optical interconnects are becoming increasingly common in supercomputing and datacom facilities – but this often means that there is a design challenge of making devices with a small enough footprint to manage the density of interconnects.4 Developments in new on-chip laser technologies over greater range of frequencies may help further improve data transfer rates by allowing for a large bandwidth of information to be sent.

Conclusions

Broadband datacom filters are integral in allowing transmission of the entire O-, C-, or L-bands and have been critical in the advancement of multiplex data streams and boosting the efficiency of modern optical networks by allowing the transmission of multiple signals down single optical cables without any loss of signal quality.

Iridian Spectral Technologies are the world’s leading supplier in datacom filters – with a number of custom solutions for optical filters.

References and Further Reading

1.     Wang, Z., Wei, G., Zhan, Y., & Sun, Y. (2017). Big data in telecommunication operators : data , platform and practices. Journal of Communications and Information Networks, 2(3), 78–91. https://doi.org/10.1007/s41650-017-0010-1
2.     Lebby, M. (2001). Optical Components and Their Role in Optical Networks. IEEE, June 2001. https://doi.org/10.1109/ECTC.2001.927765
3.     Grobe, K. (2013). Optical Wavelength-Division Multiplexing for Data Communication Networks. Handbook of Fiber Optic Data Communication: A Practical Guide to Optical Networking: Fourth Edition. Elsevier Inc. https://doi.org/10.1016/B978-0-12-401673-6.00005-2
4.     Tekin, T., Pleros, N., Pitwon, R., & Hakansson, A. (Eds.). (2016). Optical interconnects for data centers.