Flex Spectrum: What Is It?
In order to fully understand and appreciate what Flex Spectrum is and how it works, we need to jump in our time machine and go back to the year 2005. In 2005, Dense Wavelength Division Multiplexing (DWDM) and Reconfigurable Optical Add-Drop Multiplexer (ROADM) technologies were in their juvenile phase, with 32 and 40 channel systems spaced at 100 GHz.
These systems had the ability to drop all of these wavelengths at any given location throughout the DWDM/ROADM footprint and moreover, were reliable with long-life spans. Not much shortly after 100 GHz systems became available, we found the ability to shorten the channel spacing down to 50 GHz, allowing for even more wavelengths to be pushed across DWDM and ROADM systems. By combining odd and even channels across 50 GHz channel spacing, system operators could push as much as 80 or even 88 DWDM wavelengths across a system.
Unfortunately, enough bandwidth is never really enough. Exponential Internet and cloud-based traffic growth has caused this spectrum to start to fill. Moreover, most all of these systems were built for 10 Gigabit Ethernet and below speeds, requiring logistical nightmares for operators to continue to provide the necessary bandwidth.
Fortunately, coherent 100G and beyond technology has allowed these legacy DWDM systems to extend the lifeline. However, the window allotted for operators allowing continued use of these systems is coming near an end.
Ultimately the next step of DWDM and ROADM evolution is Flex Spectrum.
Fixed grid limitations
It is important to understand channel spacing on a standard "fixed grid" solution in order to appreciate the capabilities of flex spectrum. To better understand this, please reference Figure 1 below.
We see a standard 50GHz spaced system using a combination of 40G, 100G, and 400G signals. Each signal goes into its own frequency with the appropriate amount of spacing between each signal. Additionally, depending on the modulation used for each wavelength there are inefficiencies and wasted spectrum when comparing how much space a 10G uses verses a 100 or 200G wavelength.
As seen above, there is a required amount of rigid channel spacing wasted between each wavelength. Additionally, in the event of lesser bandwidth wavelengths, even more spectrum goes unutilized. This is highly inefficient and doesn't allow the full use of spectrum available.
Moreover, as bandwidth continues to grow, we are starting to use modulation and spectrum where a standard 50GHz channel spaced network simply won't allow enough of a window for 600G and beyond wavelengths to fit. Due to all of these limitations and inefficiencies, a new solution must be adopted to bridge the gap of wasteful spectrum usage and channel spacing simply not allowing for higher bandwidth waves to fit.
Flex Spectrum capabilities
Now that we understand the spectrum limitations a fixed grid solution has, let's focus on how Flex Spectrum can help. As one can probably guess, the 'flex' in Flex Spectrum alludes to that spectrum being flexible — meaning it is not confined to the standard 50 or 100GHz spacing rules as past generations.
However, despite this flexibility, guard band spacing is still necessary to avoid cross talk between wavelengths. The below diagram displays various different wavelengths being transmitted, while keeping intact the guard band spacing between services. As seen below, we can pack a Superchannel with multiple services as shown in the case of the 400G or 1Tb services. These are made up of various different sub rate services to create one Superchannel.
By doing this, the spacing between each wavelength is tighter and less spectrum is wasted. Additionally, as in previous generations the same single service is also possible as shown in the 100G service. It is still possible to create single service wavelengths, or create Superchannels to carry larger single frequency service (ex. 600G wave) or to stack multiple media channels to create one larger pipe.
Simply put, Flex Spectrum technology allows the operator to utilize more data that can be packed into the system while wasting less spectrum.
Example
Please see the below video for a real life example of how Flex Spectrum works.
Conclusion
Traditional 50GHz and 100GHz optical transport networks based on ITU grid standards has done well for past decades. However, as the industry transitions to 400G to 1 terabyte and beyond there is a need to adopt Flex Spectrum technologies.
With the adoption of greater speed interfaces and the use of Superchannels being able to support multiple subcarriers Flex Spectrum is the clear contender for optical carrier systems now and far into the future. Flex Spectrum systems allow for greater channel sizes but can additionally allow for those channels to be closely aligned to the signal being transported to increase efficiency.
Finally, no conversation regarding Flex Spectrum capabilities is complete without talking about colorless, directionless and contentionless (CDC) technologies. These capabilities will allow Flex Spectrum systems to have a greater deal of resiliency, flexibility and efficiency at the transport layer.