The greatest outrage from Mobile World Congress this week is the sheer amount of hype surrounding 5G. Hardware vendors and service providers are offering tantalizing promises of multi-gigabit wireless speeds delivered to phones and vehicles. Does anyone really believe that we'll see commercial 5G services in two years? I get the "Mobile" part means "Wireless," but it is fiber, fiber, and more fiber that will be necessary to make 5G gig speeds real.
Service providers around the world are in the process of deploying and polishing existing 4G LTE networks. Current wireless spectrum is so tight that carriers want to spill over into the already messy space of Wi-Fi in order to increase download speeds, even as they deploy ever smaller 4G cells in higher densities for better in-building and urban coverage.
Keep in mind the upper end for LTE Advanced speeds -- and this is before we go 5G crazy -- is a peak rate of 1 Gbps downlink and 500 Mbps uplink, with various trials - not commercial deployments, trials -- around the world bouncing around in the neighborhood of 250 Mbps to 300 Mbps in Europe and up to 700 or so Mbps in Asia.
But the world isn't really at LTE Advanced speed yet. OpenSignal's latest crowd-sourced LTE speed data for the second half of 2015 puts Singapore and South Korea at top of the LTE pyramid with 34 Mbps to 37 Mbps average. The U.S. has broad geographic coverage at anywhere from 70 percent to 90 percent, but download speeds are a paltry 8 Mbps to 15 Mbps.
When the smoke clears, you have to consider how U.S. carriers intend to deliver 5G. If they are lucky, they already have a decent, fiber-based network built out to support LTE and all they have to do is roll into all the towers they just finished upgrading to drop in newer, faster network equipment for more bandwidth along with new, not-yet-out-of-the-lab-or-in-production radios for faster wireless speeds.
Two years to 5G? Sure...
More realistically, carriers are going to have to get more fiber into their network for three reasons. One is sheer speed, assuming they want to deliver on the promise of multi-gigabit wireless services. SK Telecom and Ericsson demonstrated 25 Gbps download speeds while Verizon has been whispering into reporters ears about "multiple gigabits per second speeds." You can't do that over copper, bonded copper or even older fiber, because the baseline assumption I'd be making is you want to support multiple users for multiple gigabits. Give the state of the art with network speeds at anywhere from 40 to 100 Gbps, it is likely you will want more than one broadband pipe going to a distribution point to prevent congestion if you are servicing multiple customers.
More fiber will be needed for reliability. Delivering reliable gigabit broadband will mean sweeping away the mess of copper and fiber deployed over the past decade or two and installing clean sheet fiber able to be upgradable from GigE up to 100 GigE. Single point distribution of broadband might have worked if you were expecting the max load on a cell tower was around 1 Gbps; in case of a backhoe incident, you might be able to fall back to broadband copper until someone could get out and splice things. Multi-gig speeds will require more robust networks that are agile enough to route around either physical issues such as fiber breaks and hardware failure, as well as to deal with network congestion and any SDN/NFV hiccups.
Finally, more fiber will be needed because more "towers" will be needed. The radio frequencies to be used for 5G don't go very deep into buildings and structures and don't travel very far. If you thought LTE needed more towers, you're going to love 5G because carriers will need more distribution points and clever ways to route and bounce signals around to reach end-user gigabit speeds.
The first harbinger of more fiber may be found in Verizon's $1.8 billion purchase of XO. Verizon gets more fiber – something it hasn't really done a lot of with organically since shutting down new FiOS deployments a couple of years ago. Some of that fiber will no doubt get tapped for supporting further 4G and future 5G deployments.
AT&T may be in better shape, having gone gigabit and fiber crazy with threats from Google Fiber in its markets. And where AT&T doesn't have a fiber footprint, it (and Verizon) will be able to leverage gigabit fiber deployments around the country by CenturyLink, Frontier, and Windstream.
The key takeaway here is anyone that has put fiber down over the past decade should be able to find buyers in the cell tower space and by selling directly to wireless carriers, because 5G is going to need plenty of capacity.
