A black swan event is defined as “an unpredictable event that is beyond what is normally expected of a situation and has potentially severe consequences.” The phrase originates from the story that for years, people assumed that there was no such thing as a black swan because nobody had recorded seeing one – until finally someone did. It turned out black swans existed but were extremely rare, and it was hard to predict when or where someone would encounter one.
The COVID-19 pandemic in some ways has been a black swan event, and the communications sector has been no exception. One impact on our sector is that data traffic has shot up at an unprecedented rate, a result of schools closing and orders to shelter in place and work from home. Verizon, Telecom Italia, and Virgin Media are just some of the service providers reporting massive increases in data traffic. What lesson does this experience hold for the design of physical, layer 1 networks? Plenty, because this experience is one example of a fundamental truth of network design: patterns of demand in data traffic are hard to predict (other than that they will grow rapidly on average). The solution is to focus on building physical access networks that have the capacity to respond to changing demands, and the accessibility to make use of that capacity where and when it is needed.
Though the COVID-19 crisis may be a “black swan”, it does illustrate a point about the nature of data traffic: at a micro level, its growth is variable and the timing and location of demand growth is hard to predict. While what we are seeing now is a geographically distributed spike in traffic, local spikes happen all the time. Sometimes they are things we can predict, like 80,000 people deciding to gather in a relatively small location and share lots of videos on a Saturday or Sunday afternoon in the fall. But other times, they are harder. And the emerging use cases and capabilities of 5G will drive even more unpredictability.
Consider the following ways data traffic can shift dramatically in a network:
- A wireless service provider decides to build a millimeter-wave 5G network in a certain area, and now seeks backhaul/fronthaul connections every 100 meters throughout an area of a couple square miles. While we can anticipate that these types of deployments will happen, it is difficult for a wireline provider to predict exactly where a wireless operator will place their radios, how many they will place, and how many fiber connections are needed per radio.
- A new business usage pattern emerges. Consider a garage that used to hold a car repair shop which needed email, voice, and parts ordering – and is now being replaced by a custom manufacturing shop with 3D printers and other equipment that is highly automated and needs to regularly receive massive CAD files from customers.
- Consider, also, an existing large factory. At some point in the near future, they may find they now have a business case to completely overhaul their equipment with more automation and monitoring, which could generate 10-100x the data traffic.
- An ecosystem develops to support large numbers of autonomous and interconnected vehicles, but it has data backhaul and edge computing requirements that we haven’t foreseen.
We could easily predict that the aggregate data demand across a wide area of the network was going to grow at a certain percentage each year. But in each of the above cases, the demand at certain locations in the network spikes immensely, and at a rate faster, much faster, than the installation of a physical network. What is a service provider to do when planning for a physical access network, knowing that many events like the above will happen, but not where and when? Let alone events that are harder to predict?
The first thing a service provider needs to do is make sure that network has expandability: the spare capacity to accommodate additional demands. The good news is that the incremental cost of adding this capacity is low. Much of the expense of installing a new or upgrading an existing access network comes from the installation of cables: boring a path underground for new conduit, running cables from pole to pole and attaching them, etc. However, many of these costs do not vary with the size of the cable installed, and others don’t vary proportionally. Aerial installation of cable gets more difficult as the cable gets heavier, but it does not double if you double the fiber count. Boring for a new conduit, or pulling new cables through an existing conduit, also does not vary with the fiber count of the cable – as long as it fits. And new cable technologies such as AFL’s SpiderWeb® Ribbon and Wrapping Tube Cable, can fit more fibers than ever before into smaller diameter cable sheaths.
Overall, case studies indicate that doubling the number of fibers installed adds only 5-7% to the cost of installing a new fiber access network. By comparison, coming back later to add more fibers – by pulling new cables – is substantially more expensive, and could be up to 100% of the cost if the existing infrastructure needs to be precisely doubled. From an economic standpoint, then, there is a compelling argument to plan ahead for capacity.
Merely having this expandability is only one part of the equation. The network must also be flexible – meaning the service provider can readily configure it to take advantage of new opportunities. AFL addresses this challenge with modular solutions that provide a straightforward means to reconfigure the network. AFL’s SpiderWeb Ribbon is designed to simplify the task of breaking out and connecting anywhere from a single fiber to multiples of 12 fibers, to provide the flexibility to handle a range of scenarios. The ASCEND® platform is also designed to be a point of flexibility, offering a variety of modules which can be readily switched out as the network evolves. AFL also supports a range of options for PON networks, including centralized, distributed, and even cascaded splitting, to provide flexibility in network design now and into the future.
For a service provider to make use of this expandability and flexibility, though, the network also needs to be accessible. This means that it is straightforward for the service provider’s technicians or contractors to make the necessary changes to the physical network. In the example given previously, when a wireless provider comes in and installs those mmWave sites, it does the service provider no good if that high fiber count cable is solidly locked away in a conduit; it is still expensive to capture that revenue opportunity if the service provider has to go find a slack loop hundreds of meters away in order to break out a few fibers for backhaul or fronthaul.
Fortunately, AFL can solve these problems as well. AFL splice closures manage higher fiber counts in a smaller space than others, meaning they can be economically placed in smaller handholes and in smaller areas on a strand, and therefore can be closer to where they are needed. They store fiber slack internally, allowing for future splices at any closure along a cable run. They have modular splice trays which can accommodate single or ribbon fusion splices in the same tray – so if a future demand requires a single, 2-fiber, 4-fiber, or 12+ fiber connection, the technician uses the same tray. Finally, they are more easily re-enterable and re-sealable than competing closures. The AFL IDEAA® cabinet allows for quick and easy changes to subscriber connections with a modular design and pass-through capability. The AFL Titan RTD® terminal with TRIDENT® drops allows for quick and simple installations, and in conjunction with the IDEAA cabinet can also be oversized, with the ability to activate spare capacity quickl and easily.
Making use of all these capabilities requires the design and engineering skill to plan for and install a network with spare capacity that is also accessible. AFL can help here, as well. Our network solutions business has extensive experience designing capable and accessible fiber networks for some of the largest and most forward-thinking service providers in North America. Our Enterprise Services and Optical Telecom division are doing the same for indoor, campus, and venue applications.
Contact us for more information about access networks which meet today’s demands, have the expendability to meet tomorrow’s, the flexibility to adapt to changes (even black swan events), and the accessibility to make those adjustments feasible!
Written by Josh Simer, AFL Service Provider Market Manager