This blog is first in a series of 2 blogs on the evolution of trans-Atlantic cable capacity.
When we look at the current state of optical fiber transmission technologies, we may feel as if it is a never-ending journey towards greater optical transmission capacities – especially within fiber optic submarine cable systems. The past three decades have brought with them several technical breakthroughs, beginning with the introduction of optical amplification into submerged repeaters, followed by the advent of optical wavelength multiplexing and optical coherent transmission technology. Most recently, we’ve seen a significant increase in fiber count. These milestones have greatly enhanced the capability of submarine fiber optic cable system technology to meet the insatiable traffic demands of the world’s population, and it must not be taken for granted.
If we take a closer look at the historical evolution of subsea fiber optic cable capacity and what is coming in the short term, the situation undoubtedly looks less rosy. The latest cable systems have the ability to transport more capacity than their predecessors, but the new challenge for system designers is the shrinking gap towards the fundamental Shannon limit, which places an upper limit on per-fiber capacity, and the practical electrical and mechanical limitations that coincide with respecting this limit. This situation raises a simple question: Can foreseeable evolutions in cable technology handle forecasted traffic demand growth rate?
This post will focus on the evolution of the capacity of trans-Atlantic cable systems for two reasons. First, the trans-Atlantic route is traditionally where the newest cable technologies have been introduced over the past 150 years. Second, it offers superior cable capacity than the other trans-oceanic routes, due to its moderate length and strong pressure to minimize the cost per transported bit. The next post will examine whether the evolution of subsea cable technology allows to cope with bandwidth demand growth rate.
Brief Overview of Trans-Atlantic Cable Technology Evolution
The saga of trans-Atlantic cable systems started in 1858 with the short-lived deployment of a telegraphic cable system between Ireland and Newfoundland that operated for only three weeks before ultimately failing due to a number of mounting issues.
While a failed, yet still encouraging undertaking, engineers would build on it to improve the cable design and technology, eventually offering higher capacity and longer lifetimes. In 1956, the TAT-1 (Trans-Atlantic No. 1) cable system between Scotland and Newfoundland was the first trans-Atlantic telephone cable system based on coaxial cable technology, initially carrying 36 telephone channels before being upgraded to 72 channels in 1960.
The next subsea cable breakthrough occurred following two major inventions in the 1960s and early 1970s. These were the availability of compact semiconductor lasers operating at room temperature, and the use of optical fibers offering low attenuation at wavelengths where semiconductor lasers emit light. With the use of these technologies, the TAT-8 cable system entered into commercial service in 1988. TAT-8 utilized 1300 nm single mode fiber and 280 Mbits/s transmission technology. The very next trans-Atlantic cable entered into commercial service in 1991 and doubled the transmission rate of TAT-8 by operating at 560 Mbit/s per fiber in the 1500 nm optical window, where optical fiber attenuation was (and still is) minimal.
The next technological innovation occurred in 1996 with the operation of the TAT-12/13 cable systems. Both of these cables introduced optical amplification within the undersea repeaters. Unlike the previous generation of repeaters, which centered around optical-electrical-optical regeneration for reshaping the signals in the electrical domain, the TAT-12/13 repeaters employed optical amplification. Repeaters utilizing optical amplification can process multiple modulated optical carriers at the same time and are bit-rate and protocol-agnostic, allowing for upgrades with higher-speed transmission equipment when made available during the lifetime of the cable system.
A number of technical evolutions would carry the industry for almost two decades. However, it wasn’t until the mid-2010s when the industry saw the first trans-Atlantic cable systems (i.e., EXA Express and AEC-1) designed for use with optical coherent transmission technology. Although coherent transmission techniques were introduced in radio-based communication systems a few decades earlier, the ability to adapt this technology and enable its use for undersea systems was a significant achievement. It provided a boost in per-fiber capacity by a factor of approximately 10 when compared to previous generations of transmission technology.
One last characteristic to keep in mind is the total number of fibers within an undersea cable. To enable bidirectional transmission, two fibers are required, which is known as a “fiber pair.” While the number of fiber pairs within a cable system started at two and eventually grew to eight, it was not until 2021 that cable system designs could accommodate more than eight. Trans-oceanic systems today are able to support up to 24 fiber pairs (or more in the near future).
Increases in Trans-Atlantic Cable System Capacity
To assess the pace at which trans-Atlantic cable capacity has increased, we need to look at the per-fiber capacity and the cross-sectional cable capacity of past trans-Atlantic cable systems. In the case of a ring architecture (with two physically separated cables between North America and Europe), we will only consider the cross-sectional capacity of one cable. In the case of “Y” design (e.g., with one landing site in North America and two in Europe), we will consider the capacity of the trunk cable between North America and the branching unit splitting the cable into two European branches.
Figure 1 below represents the cross-sectional capacity from the TAT-8 cable system to the latest announced trans-Atlantic system – a yet to be publicly named cable from Meta that will be supplied by NEC and is planned to enter into commercial service in 2024. For the cable systems that are still in service, the thickness of the horizontal bars represents a graphical estimation of the capacity achievable with current generation Submarine Line Terminal Equipment (SLTE) transmission technology.
Figure 1: Evolution of Cross-Sectional Cable Capacity across the Atlantic Ocean
From the TAT-8 to MAREA cable systems, significant effort has been applied to increasing the per-fiber-pair capacity, with the number of fiber pairs of these cable systems comprising anywhere between three and eight, depending on the number of co-owners and their respective business models. In order to make efficient use of the limited electrical power available to the repeaters, system designers began to increase the number of fiber pairs for the post-MAREA cable systems. This resulted in 12 fiber pairs for Dunant in 2021; 16 fiber pairs for Grace Hopper and Amitié in 2022-2023; and 24 fiber pairs for Meta’s upcoming trans-Atlantic cable in 2024.
Referring back to the use of optical coherent transmission technology, a twofold impact on subsea cable system capacity unfolded. For legacy systems, new coherent SLTE technology enabled cable systems to multiply their original design capacity estimates by up to 10x. For newer cable systems (deployed since mid-2010s), a new cable system line design was adopted, which relied on an uncompensated link with highly chromatically dispersive fibers. This new line design, combined with the use of coherent SLTE transmission technology, stronger Forward Error Correction (FEC) code, denser carrier spectral packing, transmission margin optimization, and additional innovations from SLTE vendors, resulted in an original design capacity 20x higher than a pre-coherent cable system design. These advances truly paved the way for cloud computing, high-definition content, and other technologies prevalent in today’s society.
Interestingly enough, the design capacity for a given cable system is a figure that can significantly move over time, even for coherent systems. In 2015, the MAREA cable system was originally specified to support about 13 Tbit/s per fiber pair. Three years later, once MAREA was in commercial service, Infinera demonstrated a per-fiber capacity of 26.2 Tbit/s, before demonstrating 28 Tbit/s in 2021 (with some margin for commercial operation). If eliminating this margin, Infinera was able to demonstrate up to 30 Tbit/s per fiber pair!
When comparing this figure to the Shannon-limited capacity of about 42 Tbit/s for the MAREA cable system, as driven by MAREA’s large effective area fiber and short repeater spacing, the gap to the Shannon-limited capacity is about 28% (i.e., as little as 1.5 dB) for MAREA. SLTE vendors continue to improve their subsea technology found within optical transponders, but it is clear that the ability to reduce this 28% gap with the Shannon limit even further will be expensive, and that entirely closing the gap may not realistically be achievable. Unlike what the industry experienced over a decade ago, no technical breakthrough will ever allow a subsea cable system to increase its optical fiber design capacity by a factor of 10 again.
Given this realization, system designers had to start exploring other paths to achieve higher cable capacities. From 2018 onwards, instead of trying to further increase the per-fiber capacity and compensate for high-fiber nonlinear interference inherent in this approach, designers, suppliers, and owners focused on maximizing the cross-sectional cable capacity by increasing the overall fiber count within a system. Although the increases in per-fiber capacity have slowed significantly – and have actually decreased slightly when compared to the capability of the MAREA cable system, which can support up to 20 Tbit/s across the Atlantic Ocean (a standard figure based on current transponder technology) – the net result is a continuous increase in the cross-sectional cable capacity, as shown earlier in Figure 1.
Now that we have reviewed the evolution of cable capacity, our next post will build upon this knowledge to examine the implications of its growth rate, as enabled by submarine cable technology, upon the number of cable systems that need to be built across the Atlantic Ocean to cope with traffic demand.
For comments or questions, please contact us.