We're a couple years out on an unprecedented boom in broadband, with new wireless schemes, 5G, high-flying drones and constellations of satellites covering us in a mesh of data – if the media hype is to be believed.  But all of these “new” ideas are the latest iterations of older ones that didn't make it past basic R&D.  Put your faith in fiber, in my opinion, and be skeptical of everything else until there's a sustainable business going.

Let's start with a skeptical examination of the latest “new” technology, fixed wireless.  AT&T, Google and Verizon are all testing fixed wireless schemes, with the two cellular carriers relying on 4G LTE and to-be-determined 5G technology for faster speeds with less fiber infrastructure deployments.   Fixed wireless services require a cell tower with fiber capacity to deliver services to subscribers, typically with an external antenna for a subscriber.  No fuss, no muss, no trenching. Sounds great!

But it all depends on the radio frequencies involved and the load around the cell tower.  One doesn't typically position cell towers and microcells in low-density areas, due to demand for mobile services.  If service is being delivered in higher and/or 5G licensed frequencies, line-of-sight is needed at some point between the service provider and subscriber – easy enough in rural environments if you have a tower, but denser urban environments and cities become a pain. 

Service using 5G technology gets into bigger hand waving, since available frequencies will vary from country to country and geography to geography depending on licensing.   Because 5G requires line-of-site, any service provider playing in those frequencies will need to deploy more cells – more than the “thickening” required for 4G urban deployments – and those cells will have to have fiber to provide sustainable speeds akin to, well, fiber.  There are some interesting tricks with mesh networking to extend coverage without the use of fiber, but RF bandwidth and antennas used for networking deducts from available resources to serve customers.

Facebook is testing long-duration unmanned aircraft, drones if you prefer.   Each drone will stay aloft at 60,000 feet for weeks at a time, powered by solar energy and connecting to other drones and the ground via laser, with end-user connectivity delivered via laser beam.  But Facebook has many months of work to build and demonstrate a single production-quality end-to-end constellation, delving into aeronautical engineering territory that only NASA and the Department of Defense have explored.  The economics might not work.

Large scale satellite networks using several hundred to several thousand satellites in low Earth orbit are all the rage, with players from Boeing to SpaceX having filed plans with the FCC to reserve spectrum and launch test satellites. OneWeb appears the farthest along, with financial backing from Bharti, Coca Cola, Intelsat and contributions from Airbus, Qualcomm and Virgin Galactic  “Cloud” satellite networks offer low latency and coverage of the entire Earth's surface, but nobody's built one with a couple hundred satellite networked together. The Iridium network is the best working example today, with around 66 satellites and spares put up in the 1990s and delivering voice and lower-speed data.

OneWeb, in comparison, plans to put at least 648 satellites into orbit starting in 2019-2020, with ground users getting speeds of 50 Mbps per second or faster.   Initially, OneWeb made a lot of noise about delivering broadband everywhere on the planet, but has since shifted to pitch the ability of the network to serve businesses and extend cellular phone coverage.   Good intentions to deliver broadband to the world are one thing, but business customers will be necessary to pay the bills and keep the lights on.

How many cloud satellite networks can the market support? Nobody knows. OneWeb has the daunting task to demonstrate that a single cloud satellite network can be a successful business.  LEOSat believes there's room for a cloud network tailored to the high-end enterprise market, delivering low latency and gigabit speeds for corporate and government customers, with performance equal to or better than fiber. 

Many cloud satellite network proposals are also facing a battle with terrestrial 5G networks for radio spectrum, with satellite builders trying to stake out spectrum claims before some 5G providers can get their own equipment running.

There is no clear winner in the rush to new networks, except for fiber.  Broadband delivered by fiber doesn't have to worry about radio interference and can easily carry hundreds of gigabits to terabits of traffic per second on a single cable. Fiber has outlasted challenges from other schemes, due in part to its maturity and relative simplicity.