Digitization and data growth show no sign of slowing down, meaning you need to keep finding new ways to deliver massive network capacity at the lowest cost per bit. 400G is a solution, but how ready is it and what benefits might telcos and mobile companies get from it?
400Gbps wavelengths have been around since 2017, but there hasn’t been any real demand until more recently. Traffic across networks has historically been unpredictable, but recent increased homeworking and remote access to enterprise, education and entertainment apps has driven a massive increase in data usage.
Overall mobile data traffic is forecast by the ITU to grow at an annual rate of around 55% from 2020 to 2030, reaching 607 exabytes in 2025. Mobile video is forecast to account for 79% of all mobile data traffic by 2027. And the global cloud storage market is forecast to grow at a CAGR of almost 25% from now to 2026, as the volume of data being uploaded to enterprise servers continues to rise rapidly. Consider that around 60 million photos are uploaded daily by Instagram users around the world, and that gives a flavor of what is to come.
This ramping up of data-hungry apps and need for storage could see 400G become critical for Content Providers, Internet Service Providers (ISP), telcos and mobile operators to enable traffic in more efficient and effective ways.
To begin with, it is worth establishing that 400G capacity is not in itself new – network managers today already know how to combine several 100G ports into a 400Gbps or higher throughput – but it can make ramping up capacity more manageable and drive benefits for your business.
400G over a single wavelength can make it easier for fiber optic networks to handle the ever-growing burden of increasing data volumes. It’s a future-proofing technology that can address current and future bandwidth needs while reducing the cost per transported bit.
Using individual 400G ports rather than a combination of 100G ports can help content providers, ISP and their large customers enjoy greater flexibility to support all that increased traffic, without having to add more ports. It speeds things up at a practical level too, since initial delivery will be faster for end customers as there will be only one interface to install and test rather than four.
In terms of costs, a 400G card is not much more expensive currently than four separate 100G cards. No saving or additional expense is currently predicted on that front. However, in terms of cross-connect costs, savings will be noticeable as only one cross-connect will be required instead of four.
Furthermore, by being able to deliver four times higher bandwidth per optical transceiver, telcos can drive a lower cost per bit by delivering the same bandwidth in less physical space. 400G circuits are also easier to manage with fewer overall ports.
What’s more, 400G is better-suited to supporting new and emerging use cases moving forward. 400G can give telcos and mobile companies the power, efficiency, and density required for 5G and other fast-growing applications such as virtual reality and augmented reality (VR/AR) and 4K mobile video streaming.
For enterprise customers operating an on-demand model, a 400G port enables them to start small and scale up as needed, with a pay-as-you-use model. As explained earlier, there may also be savings on the time required to deploy the initial connection.
The other side of the coin is that having a single port can create a single point of failure (SPOF). The need for redundancy and expectations on availability must be taken into account before moving to 400G ports.
400G has an impact on port density in the data center and can enable practical and financial benefits. IP routers today are at capacity, as arrays of 100G ports seek to service all that demand for traffic. However, 400G ports mean being able to deploy higher densities of ports per router, allowing for growth and servicing more data without the need to drastically increase the number of routers in place. It means you get more out of your existing hardware and don’t need to spend on more routers or more space in the data center. It also means savings on maintenance.
It is likely that 400G cards will not consume as much energy as using four individual 100G cards. Orange experts are yet to confirm this, but it should hold. What is not definite at this stage is how much “energy consumption by card” will increase or, put another way, by how much the Watt per Gbps ratio will decrease.
In terms of standards, this also remains to be confirmed. At present, there are a few varieties of 400G, including 400G LR4, 400G LR8, and 400G FR4, and the optical signal can be carried as one wavelength or several. 400G FR4 for example is adapted to distances up to 2km, and is 30% to 40% cheaper than 400G LR4, and could therefore be adapted for intra-site connections. Standards in optical modules still vary by range, and prices remain unconfirmed.
In terms of transmission, 400G can help operators further optimize existing fiber networks, meaning no need for spending on engineering work to lay more fiber cable. Spectrum efficiency is improved noticeably as well: at the end of 2021, Orange and Nokia engineers validated a first field test that confirmed Orange’s network capability to increase its optical capacity to support end-to-end 400Gbps services across its network.. This represents an increase in spectral efficiency of 50% versus previous iterations of Orange’s long-distance network.
Another decision to be made is about distance: as speed increases, the distance the signal can travel drops. So for signals to travel the same distance, operators would need to add more transponders to regenerate the signal, which means additional cost. So it’s worth acknowledging that only the routes that are expecting to see the highest increase in traffic should be upgraded. Intra-FLAP* routes however are not affected by this need for more equipment as distances between them remain relatively short.
| *short for Frankfurt, London, Amsterdam and Paris, the cities in Europe where the largest data centers are located. By extension, the acronym FLAP may also include Marseille. |
The rapid evolution of the technology means that there is an impact on transmission equipment. It presents the possibility of replacing grey interfaces on the router side with colored interfaces that are able to replace transceivers (TRX) and connect directly to the ROADM. Colored interfaces are now both a smaller form factor (QSFP-DD) and lower cost than they were previously. That said, these colored interfaces are still less efficient than traditional TRXs in terms of distance the signal is able to travel. As a result, on long routes that require 400G, using colored interfaces might mean needing to spend more on amplification and regeneration equipment.
Furthermore, there is a relatively new feature available on the transmission layer that could help telecom operators monitor how much bandwidth existing customers actually use on their 100G circuits. Depending on whether the largest customers use most or little of their available bandwidth, and how that consumption changes over time, telcos will be able to evaluate whether or not these customers might benefit from end-to-end 400G, and plan investments more accurately.
This is a key question, and as of this moment, the answer is very much “it depends”. Demand from end-users is there for much greater traffic across many networks, and the technology seems ready to go. Our expert engineers have evaluated 400G and see it as a viable route forward for telecoms companies and mobile operators who need to deliver much greater capacity.
For ISPs and Telcos in general, it is very much worth considering 400G from the router perspective because major savings can be had in terms of hardware spend and data center space, while supporting much greater traffic. But as mentioned, it isn’t worth investing if your traffic increase isn’t going to be that big – there’s no financial value in having 400G cards sat empty and unused.
It is also worth remembering that when 100G arrived, it followed soon after the availability of 40G, so many operators made the move straight from 10G to 100G and bypassed 40G completely. That looks less feasible at the moment since 1Tbps remains a long way off and may not in fact be available until at least 2024, according to Orange network engineering experts. So if your network has the need, and the demand is there, it is probably worthwhile investing in 400G. However, if you are only using smaller points of presence (PoP) where traffic growth is likely to be less, it may not be worth the investment right now.
From the enterprise or content provider perspective, 400G could enable greater flexibility and growth, as there would be less need for new interfaces with the ISP, although relying on a single port could increase the risk of service failure in the event of an incident. Some costs savings are to be expected, at least on the cross-connect front.
Moving forward, with overall data consumption expected to keep on rising rapidly, 400G will increase in use. Given the ever-increasing data requirements of so many areas of work and home life, including IoT use cases and 5G network infrastructure, telcos and mobile operators may need to make the move to 400G to enjoy greater spectrum efficiencies and cost savings while delivering great QoS for data-intensive applications.
