Chattanooga’s City-Owned ISP Pushes 25 Gbps Broadband
from the so-this-is-what-competition-looks-like dept
Chattanooga, Tennessee is one of a growing number of U.S. cities to build its own broadband network. The ISP, tacked on to the city’s existing EPB electricity utility, has routinely delivered speeds faster and more affordable than the services provided by regional utilities like Comcast.
While community-run broadband is portrayed as boondoggle socialism by telecom monopolies and the folks paid to defend them, the ISP has generally been a success story. It routinely provides speeds upwards of 10 Gbps to locals, has repeatedly shown the highest consumer satisfaction ratings in the country for broadband, and drove significant economic value to the city.
PC Magazine also recently ranked Chattanooga as one of the best work at home cities in 2021 thanks to its city-owned broadband network.
The city’s now pursuing another first: the first ISP in the country to deliver residential broadband speeds as high as 25 Gbps. It’s not cheap ($1,500 per month for residential, $12,500 a month for commercial users), and it’s a speed few people truly need, but it again demonstrates how it’s community broadband ISPs pushing competitive barriers in the U.S. telecom space:
“We are once again breaking the typical approach for internet service providers by proactively upgrading to the latest technologies in anticipation of future needs. Our goal is to enable new frontiers for technical innovation and job creation for our customers to the benefit of our whole community.”
Studies have routinely shown that community broadband services provide faster, less expensive broadband at more transparent rates than most private offerings. They’re not a mystical panacea, but they can often force stodgy old regional monopolies to improve service. And, contrary to the industry narrative that they’re an inherent boondoggle, there are plenty of very successful efforts.
If you recall, Comcast unsuccessfully tried to sue Chattanooga’s effort out of existence, and lobbied state lawmakers to pass a nonsensical state law banning the operation from expanding. Comcast ultimately buckled and was forced to provide faster broadband at more affordable rates, but not until after it had used a myriad of dirty tricks and lawsuits to try and kill off the effort.
Filed Under: 25 gig, broadband, chattanooga, community broadband, digital divide, gigabit, high speed internet, municipal broadband
Comments on “Chattanooga’s City-Owned ISP Pushes 25 Gbps Broadband”
It’s not a speed that most people could even use. 1 gigabit/s is above the routing capacity of many current routers designed for residential use—and with even this speed being rare, manufacturers don’t tend to list expected throughput on the box. Dual-core ARM-based systems typically work well.
At 10 gig, you’re into speciality/business devices or just AMD/Intel PCs (possibly with separate wi-fi access points). At 25 gig, you may have trouble finding servers that can keep up with you. Init7 in Switzerland sells gigabit, 10-gig, and 25-gig service for the same price (except for the setup fee), about $80/month; so of course some people have signed up “just for fun” and noticed these things.
Re:
It’s not a speed that most people could even use.
True, but with one important addendum: ‘Yet’.
It wasn’t too long ago that measuring speeds in the megabit range would have been seen as both blindingly fast and completely unnecessary so while speeds in the gig range might be overkill currently it’s almost certainly only a matter of time until someone finds a use for a speed that high.
Re: Re:
That may be a bit of an exaggeration. In the early days of the consumer internet, “everyone” knew of ISDN and T1 and hoped to eventually have such speeds. And when I got 3-meg cable, it was obvious that it wouldn’t end there—a 700MB movie still took like half an hour to download, OS updates were kind of slow, etc. With LANs being 10 Mbit/s, and not difficult to saturate, the internet link was a chokepoint. It was the T3 and OC-12 that would’ve been complete overkill, with our computers and local networks unable to keep up.
Anyway, does anyone else think the FCC’s broadband definition of 25 Mbit/s download and 3 Mbit/s upload is embarrassingly low when 10000/10000 is affordable and 25000/25000 is available? Some form of open-access fiber to every house should be the minimum standard, and we knew that decades ago.
Re: It's partly about the backbone connectivity
In the pandemic lockdown, our accordion ensemble used Jamulus for continuing rehearsals. It’s not much of a bandwidth problem since it uses the Opus codec (but of course raw data could stream with less latency) but latencies are very much determining what kind of music you can still rehearse successfully. Living in the same town does not mean sharing the same provider, so you end up renting a server in network vicinity to an Internet handover point, probably 100mi away (this is densely populated Germany).
The luckier get about 30ms roundtrip to server and back until their own sound arrives in their ears in their personal mix. The unluckier ones get 80ms or more, making it hard to do their share in keeping robust speed.
Reliable low latency across your town (and no long detours to distant backbones) would certainly have helped.
Now that’s audio-only. How about being able to do the same with visuals? Certainly would be nice seeing the conductor, or a player-conductor.
Again, you want low-low-low latency but now at video-feed speed, with a codec that doesn’t get to be highly efficient because there isn’t the time for that.
25Gbps continuous streaming speed? That would seem to be less important than being able to work with low latency. Unless you are operating a web crawler…
Re: Re: Serious video ciompression
I would solve the video conductor problem by reducing the video transmission to basically a stick figure; maybe we’ll get her face in there at a bit of latency in case she grimaces at a missed note.
But it’s a good problem: Can our favorite symphony orchestra effectively rehearse over the internet in real time?
Re: Re: Re:
With sufficiently high-bandwidth networks, the difference between “stick figure” and “realistic image” shouldn’t matter.
I’ve seen various numbers for tolerable latency for live band practice, ranging from about 10 ms to 30 ms (around 5000 km distance). Consistency matters too. It would be good to have some real (e.g. FCC) standards on this.
Opus latency defaults to 20 ms but can be set as low as 2.5 ms. If one’s concerned enough about that to consider raw PCM, one will also want to carefully check soundcard latency, which can be a similar amount by default. I guess everyone involved should use the same ISP, to avoid the additional delay of external routing. If your “favorite symphony orchestra” is well funded, it wouldn’t be absurd for them to consider building their own network on dark fiber—yet another reason every home should have several dark fibers available for lease.
Re: Re: Re:2
When throwing around latency numbers, remember that sound travels aobut 1 ft/ms in air and that roundtrip latencies correspond to double distance. If 30ms is too much roundtrip latency, then 5 yards would be an intolerable distance between band members.
How do organ players manage when their own instrument is larger than that?
Yes, even small latencies require speed discipline: you cannot afford voicing only once you hear someone else. Then even the smallest latency will lead to a performance speed crawldown. People manage that even without the Internet in between.
With good microphoning and a good soundcard, your monitoring mix is very direct, giving you the impression of the others being right next to you while you need the musical reflexes for them being 5 or 10 yards or more away. It takes practice.
A near-instantaneous Internet in town would make quite a difference.
Re: Re: Re:3
You mention that people can get used to in-person latency. Note that unlike most internet latency, that’s 100% consistent. (Cable modems, when the local segment is overloaded, can be especially bad—I’ve seen latency occasionally spike to 5 or 10 times its usual value, on a per-packet basis.) And don’t forget that the musicians can see each other (and/or the conductor), with negligible latency, at in-person sessions.
Re: Re: Re:3 Actual experience?
I’m idly wondering if anyone here has had actual experience with these delays and positioning effects, whether with band, orchestra, or choir.
Noting that way back when, Windows 98 was notable for 10 and 100ms “absences” when trying to run “real time” code, so I see getting packets onto and off of the network as a potential obstacle to realizing effective remote musical practice.
Re: Re: Re:4
Yes, a few musicians or a small choir can be in sync by just using auditory cues from the others. It becomes a problem when the group is large enough that the auditory cues become muddled and mistimed due to distance which makes it very hard to latch onto one specific cue. Depending on the situation you then use visual cues (conductor, choir leader etc) or a metronome.
It should be noted that it doesn’t really matter that singers in a choir is slightly out of sync due to the distance to a microphone as long as they have the visual cues, because they are still human and they will not have the same timing anyway when singing. Same for large bands.
Re:
But if you wanted to drop a CDN someplace with huge bandwidth that wasn’t going to cost an arm and a leg you’d see all sorts of them parked aroudn town.
Re:
most of the routers aimed at gamers have 2.5 gig ports in them, and “port teaming” for even higher speeds.
unless you are talking in the sub 60 dollar space(which is considered budget) damn near every router on the market is designed for 1g/s, most differences are how well it does at handling traffic depending on how many devices you have connected at a time.
as for listed throughput, netgear, tp-link, asus, linksys, and d-link(to name the first few that come to mind) all list throughput on their boxes, with the exception of again, sub $60 equipment by linksys, and every Belkin device I’ve ever seen 😛
Re: Re:
If you spend $60 and don’t get gigabit transfer you payed too much.
A $25 NetLnk knockoff can handle 1200mbps.
Welcome to Streamer Town.
Hard to tempt someone with bread and water when they've had a feast
Another fine example of why the incumbent companies are so desperate to kill off or cripple community broadband, because when it’s allowed to flourish it not only provides serious competition they also show the sort of offerings those companies could be making available if they bothered to and weren’t focused on offering the least they can get away with at the highest price they can think of.
Pinky, are you thinking what I’m thinking? Maybe Big Telecom tries to sue municipal broadband out of existence because it owns the cable companies, and if their captive markets have shitty Internet speeds, maybe there’ll be less cord cutting in favor of streaming.
On par, ‘I guess.
Comcast offers a 10Gb/1Gb business service here that’s between $900-$1200 per month.
Residential Gigabit is around $80 (1100/25)
Probably because we have overlapping national (4 options) and independent (dozens of choices). No community choices here though. And sadly most choices have cap limits. Semi-unlimited is another $25-$50 on all but the 800Mbs and higher options. A bit of a joke when 500Mbs at $55 plus unlimited at $35 is MORE than the top plan for Comcast
The article is missing the cost of a realistic speed for comparison here.
That said, nothing beats competition. One more choice for the Train people is great.
I want it!
I would like to use something like in place where I live eve (I currently have 750 + 500 Mbit/s from 2 different ISPs(with Mikrotik as Dual-WAN router) and I can get 2 other ISPs. No way to get something more than > 1 Gbit/s).
I’m not from USA but IF/WHEN I decide to emigrate to USA – I knew place where I want to live!