Connecting the Unconnected


In Episode #58 of the podcast, Tom Evslin joins Richard for a discussion and demonstration of Starlink in Vermont. Tom signed up for the Starlink beta and he uses it for our Zoom session.

Tom is a genuine Internet luminary: he pioneered flat rate Internet pricing when he ran the AT&T WorldNet dialup Internet service as well as the use of the Internet to transfer voice. A company he founded, ITXC, was the first major supplier of international VoIP transport across the Internet.

Tom was also a guest on the first High Tech Forum video podcast in 2017. He signed up for the Starlink “better than nothing” beta test and was pleasantly surprised. He joined the podcast via Starlink so this is a great demonstration of the Starlink Low Earth Orbit satellite-based broadband service.

Why Starlink?

Tom lives in rural Vermont where broadband options are severely limited, so he signed up with Starlink to see whether it would be an improvement over existing alternatives. Vermont has a program underway to subsidize rural broadband in unserved areas, but the program is unlikely to provide meaningful alternatives for five years, if ever.

The Starlink FAQs on his blog, Fractals of Change, describe the state of Starlink during the beta. Since then, speed has increased to the 300 Mbps down/30 Mbps up range, with latency in the neighborhood of 45 milliseconds.

This is fast enough for every common Internet application, but it falls short of the symmetrical broadband desired by some broadband policy figures in Vermont and Washington DC today. We’re still waiting for a list of applications that will be liberated by very high speed symmetric service.

The Starlink Technology is Awesome

Tom’s first impression was surprise that there were no instructions in the box when he got his Starlink dish. But he quickly found that he didn’t need them.

When you plug in the dish, it automatically aligns itself to the Starlink constellation and before you know it you’re in business. You can see the real time map we discussed by clicking here.

Tom found the handoffs to be utterly seamless. His initial siting of the dish was partially obstructed, so he experienced a loss of connection for a few seconds every hour. He corrected that by relocating the dish.

Data Centers in Space

The most fun part of the podcast is Tom’s reflections about data centers in space. This is a takeoff on the fact that Starlink has taken to locating ground stations at data centers and building satellites that talk to each other.

That saves money and also provides better performance. I can’t summarize this part effectively, so just listen for it at 26:25 or so.

Starlink also has potential as an alternate backhaul for disaster recovery purposes. The satellites are solar powered and effectively immune to earthly disasters such as hurricanes and wildfires.

Won’t it be ironic when LEO satellite constellations are providing backhaul to FTTH access networks?

The Greatest Dangers are Political

I doubt that LEO will ever completely replace fiber backhaul and mobile networks because of capacity constraints, but it’s an intriguing piece of the puzzle that has to be solved for universal broadband to become reality.

Perhaps their greatest threats are political. We’ve already seen that terrestrial networks with political exposure can be hampered and delayed by political considerations.

5G networks need licenses for small cell siting, and many municipalities are reluctant to issue them when 5G competes with muni networks. Cities can hide behind phony health concerns when dragging their feet, but they have additional motives to delay the 5G rollout.

Vermont Wants to Protect Government Networks

At 33:15 we get into the political issues that came to the fore when Broadband Equity NOW! – the organization Tom and his wife Mary formed – petitioned the Vermont legislature to use Starlink and other existing alternatives for immediate broadband needs. Vermont’s reaction was refusal because it doesn’t want competition for its government-owned networks.

This harkens back to the themes in our last podcast, in which John Horrigan made it clear that the greatest barrier to broadband adoption is lack of information about available networks and subsidy programs.

The Digital Divide is often sold to us by politicians as a civil engineering problem, but the reality is that it’s a social problem driven by lack of interest, income, and information. For every person unable to buy high quality broadband at any price, there are three or more who can but don’t because they don’t know why or how to get connected and where to find help paying the bill.

Understanding the Requirements

Policy makers would do well to learn one of the fundamental insights of computer engineering: a successful product development cycle begins with a good understanding of the customer’s requirements. A failed project usually begins with a fantasy about a cool solution into which a user need must be shoehorned.

The failed project scenario is all we hear from Washington and many of the states today. The iron seems to be hot, so long time advocates of government created fiber networks are eager to brand all the livestock in sight with fiber.

When we begin with the requirements we quickly find that there are many ways to satisfy them. At this point it’s more prudent to continue to rely on innovation to meet needs rather than declare one and only one technology the permanent victor.

This was a fun podcast to do, so it should be fun to watch.