Tell that to the people on gigabit GPON in Grant, Douglas, Chelan, Benton and other counties in WA state right now. The county PUD electrical grid operator builds the fiber and splitters, and sells L2 transport services to independent ISPs. The ISPs do the layer 3 IP part, customer support, billing, etc. It works great.
Do you know why those counties built their last mile fiber networks? Because the incumbent telco (centurylink or frontier, for the most part) and cable (comcast or charter) private entities in the same service territory were showing no signs of building anything better than 20 Mbps down x 2.5 Mbps up.
On the other hand, Tacoma Power couldn't make the economics of their own-brand broadband Click Network work, [1] so they cut their losses to some extent by transferring operations to a private company, Rainier Connect. [2] Rainier Connect's Tacoma system is still owned by the city of Tacoma, though, and hopefully there will not be further moves toward complete privatization.
Click was built as a coaxial copper last mile network - not comparable in performance to any new build fiber based efforts in the years 2016 and later. In industry terms, it was best charitably described as "HFC" - hybrid fiber/coax, meaning they would have some small amount of fiber to the network node locations of the CMTS, but the actual houses were all served over copper.
I think we both agree on this, but in case not: I don't think Click's uncompetitiveness was due to inherent limitations of hybrid fiber/coax. These days, cable TV companies routinely sell gigabit or near-gigabit download speeds to their HFC-network customers. Yet Click doesn't sell more than 100 Mbps even now. Click's financial problems may or may not have started because they stopped being competitive - perhaps the causation went the other way - but that failure to change certainly didn't help.
For instance, I just picked a random house for sale in Tacoma and tried pricing Internet for new customers from Xfinity (Comcast) and Rainier Connect (former Click). Xfinity offered 1 Gigabit (fine print says 940 Mbps) downloads for $84.99/month on a 1-year contract, increasing to $100 after the contract term. Rainier Connect's fastest offering was only 100 Mbps and cost $104.95/month - pay more to get one-tenth the speed.
(These rates aren't exactly comparable, though I think they are still roughly so. In particular, Xfinity has a 1.2 TB monthly data limit, while it's not clear whether former-Click customers at Rainier have one. Click did impose a < 1 TB monthly data cap. [1])
The most silly thing about this exercise is that if you look at Rainier's rates in Puyallup [2], which is not a former Click Network location, you see that they definitely sell gigabit to some customers for only $80 - but apparently not in Tacoma.
> For instance, I just picked a random house for sale in Tacoma and tried pricing Internet for new customers from Xfinity (Comcast) and Rainier Connect (former Click). Xfinity offered 1 Gigabit (fine print says 940 Mbps) downloads for $84.99/month on a 1-year contract, increasing to $100 after the contract term. Rainier Connect's fastest offering was only 100 Mbps and cost $104.95/month - pay more to get one-tenth the speed
And it does not mention upload bandwidth anywhere because it will be a tiny amount split between tons of houses and the latency will be terrible compared to domestic fiber. Not to mention Comcast is probably over provisioning by a ton and selling you “1,000” Mbps “burst” speeds for 5 min or something while they slow everyone else in the neighborhood down.
I’ve learned from plenty of experience to never trust non fiber to the home marketing numbers. Coax is obviously a much inferior technology.
> These days, cable TV companies routinely sell gigabit or near-gigabit download speeds to their HFC-network customers.
but ultimately coax is polishing the brass doorknobs on the titanic. gigabit down, yes, but at what contention ratio? and how many upstream channels are bonded for a reasonable upstream speed? It's not uncommon to see something like a comcast setup that is 800 Mbps down, and 16 Mbps upstream. How much traffic throughput can a given section of coax/DOCSIS3 network support if a large percentage of people actually start using their upstream for more than a tiny percentage of the time?
in terms of future upgrade path and ultimate life span of the outside plant cable network, singlemode fiber to the home wins every time...
one theory as to why the speed offerings from click/rainier connect are so poor is that the coax plant has not been subdivided into small enough sections to actually support anything like gigabit to the consumer. now this would require going through and re-engineering it, breaking parts of the network apart physically, and installing new small CMTS in neighborhood aggregation points.
As a resident of Chelan county, I can attest, muni broadband is a great idea! It has allowed me to live the lifestyle I choose, where I choose. I’ve had <$80/mnth 1G fiber for years now.
Don't private ISPs behave this way because competition with them was outright banned? I remember most cable companies agreed to do all the heavy digging and cabling only in exchange for being exclusive service providers in the area.
So the ISPs may formally be private, but market forces are absent, with predictable consequences. A municipal network breaks this monopoly, adds competition, so, it's good.
They behave this way because the restricted nature of the physical plant of last mile access makes them a duopoly. In some of those areas of centurylink and comcast, neither one wanted to flinch first and be the first to spend many millions replacing their copper plant with fiber. Since they were both equally aware that the consumers in the area had available service options from only two companies, and exactly what the prices/packages/speeds of those services were...
In some cases where the copper phone lines are degraded or a long loop length, it's worse than a duopoly. It's a monopoly since the service options might be 5 Mbps down x 768 kbps on ADSL2+, or 200-300 down and 15 Mbps up on comcast or charter DOCSIS3.
That can be the case depending on what the population density looks like and how much of the infrastructure has to be aerial vs. underground. Doing a fully independent overbuild for GPON FTTH to individual houses can have a cost per house of $4000, which results in unacceptably long ROIs for fully private investment. Private investors who fund startup ISPs will go elsewhere.
It can also be complicated and costly as a new-entrant 3rd provider to force the local entities controlling the wood utility poles to permit you to install infrastructure through entire neighborhoods and cities.
$4k amortized over a few years is very reasonable for a utility every house will buy. The problem is that Comcast already exists and will undercut you with an very inferior product, but the general populace is not educated enough to differentiate fiber vs coax, so Comcast will drop their prices to whatever they have to do the fiber won’t get purchased.
Which bring us back to why it works for a PUD or municipality, which has the wherewithal to amortize things like that over 20 years from funding on municipal bonds, and can compel every homeowner in an area to pay for it.
The existing duopoly have deep pockets and could protect their duopoly by also laying fiber anywhere that a third company tried to start up. The new entrant would not be able to maintain a profitable competitive advantage in any market.
Comcast has upgraded its service in the cities where Google Fiber has been rolled out.
Tell that to the people on gigabit GPON in Grant, Douglas, Chelan, Benton and other counties in WA state right now. The county PUD electrical grid operator builds the fiber and splitters, and sells L2 transport services to independent ISPs. The ISPs do the layer 3 IP part, customer support, billing, etc. It works great.
Do you know why those counties built their last mile fiber networks? Because the incumbent telco (centurylink or frontier, for the most part) and cable (comcast or charter) private entities in the same service territory were showing no signs of building anything better than 20 Mbps down x 2.5 Mbps up.