translating things, building chill software for my friends, playing ttrpgs, making procedural vector art, learning piano, writing unhinged Utena fanfics, and just vibing



cathoderaydude
@cathoderaydude

I've made a lot of posts about POTS (pots posts) today, and I feel I should offer a counterpoint to all my negativity about VoIP, because - genuinely - it has pragmatic advantages.


I am generally not in favor of decentralized systems, for myriad reasons I won't go into, and the phone system is a great example of one that works really, really well because it's centralized. It's good because it's a single system, because it's not a bunch of loosely-coupled, independent networks with "compatible interfaces."

Without getting into it: The US government forced the phone companies at gunpoint to behave more like the internet on two occasions, and both made service much worse for no gain, and pushed scammy practices through the stratosphere by putting enormous pressure on tons of people to produce artificial profits.

However, there is one aspect of the phone system's centralization that was always bad, and that is: Outside Plant. That's the ancient telco term for "wires, the actual wires."

I do not wish to recite the entire history of the telco breakup here, but to summarize: A hundred years ago, AT&T began burying Wires to build a phone network. Eventually they began burying massive cables full of many wires. Consider: If you have an office tower with 300 units in it, and each one has 10 phone lines, you need 3000 pairs - or 6,000 individual wires - to serve that building.

Find attached above a picture of (what I believe to be) a 3600-pair cable. Yes, that's 7200 wires, and that's what you'd need to supply that office tower. I also included a picture of what some of the cable maintenance look like. All those white rectangles are splice blocks that join together hundreds of pairs of wire, all of which must be punched down by hand.

The amount of work that goes into maintaining this stuff is obscene. It always has been, and that's not necessarily a problem. Some of these cables were buried nearly a century ago and they still work. If it costs a quarter million in labor and permits to replace one, it's still going to have paid for itself.

But what happens if it partially fails? If the pair feeding your office in the tower goes bad... well, you could be in a pickle. The phone company is not going to dig up a 3600 pair cable and replace it in order to fix one pair.

Now, I don't know how the math here works. There IS some point (or there was, until recently) where enough pairs have failed that they will bury a new cable, and in the meantime, any given trunk to a particular building should have enough spare capacity that if your pair fails, they can move you to another, and another, until they find a good one.

But sometimes that doesn't work. I can best explain this in the context of internet service, my area of greatest experience. Namely, with ADSL, a notoriously marginal technology.

ADSL was very much designed to work over Lines Of Unknown Provenance. You can put it on brand new, just-installed wires or shitty, century-old copper, and it'll try its very best to make the most of it, which sometimes... ain't that hot.

It wasn't uncommon, in the 2000s, for ADSL to be the only option available in many places, and many people couldn't get anything faster than 384kbps. That's, like, three kilobytes a second. But if the modem tries to train up, tones out the line, and can't get reliable data through any faster than 386kbps, then that's what you get.

And if you were a residential customer... that was it. Game over. Your house had one pair going to it, maybe two if you were lucky, and if neither was any good, or if you were too far from the local telco office for a clean signal to get through, then you were stuck downloading files at 3 kilobytes per second in 2009. There was no fix. None.

Making a VERY long story short, there was little or no competition as far as "wires going to customer's houses and buildings." When AT&T was the only phone company in the country, they owned every single telephone wire buried anywhere in the nation. Even after AT&T got broken up into smaller telcos and the government legislatively forced competitors into existence, they still owned all the wires. This is still true today: in any given region, a single company - called an ILEC - that used to be part of AT&T owns every single buried phone wire.

So, while there were many DSL providers in 2009, switching to a new one didn't do any good. Sure, you could switch from AT&T to Earthlink - and they would simply place an order for the pair of wires already entering your house. AT&T would go to the central office and move the end of that pair from their equipment over to Earthlink's equipment, but - same wires, same problem.

You couldn't get new wires brought to your house, because the same laws that required the ILECs to let smaller companies "lease" the wires that they'd already buried (so every ISP didn't have to independently trench and bury wire to every single house they served) also made it economically pointless to replace them. What's the point in replacing a pair of wires that some other company is going to profit from, while you get nothing more than a tiny legislative stipend?

So, the wireline infrastructure in the country has been rotting for a century, and due to a complicated legal and economic situation, it hasn't made sense to fix it for many decades.

Now - who will dig a trench to your house and bury new wire? Comcast!

Cable companies are REALLY good at getting you nice, fresh, clean copper. Their infrastructure is half a century newer, the cables are much more robust, and it's cheaper and simpler and offers a MUCH better ROI to bury new cable - plus, they're actually in competition. The ILECs are old, dying giants who don't really care if they win your service contract or not. The cable companies are hungry to steal you away from other vendors.

In 2009, it often made a LOT more sense to try for cable internet, even if you had to pay for trenching yourself - the cablecos would often do it! The telcos would not. I NEVER, not once, heard of someone with bad POTS wiring getting it replaced. As far as I know, it's impossible, as in, no amount of money or prayer will make it happen. But I've spoken to people who said "oh yeah I paid Comcast $3000 and they buried a brand new cable right to my house."

Then, as the 2010s unfolded, fiber hit the scene. It wasn't too easy to get at first, but it's becoming more and more common for ISPs to aggressively offer the stuff, at their own personal cost. That is to say, they're acting like the telcos did back in the halcyon days. They consider the installation of new fiber to be an investment in infrastructure; they're buying a customer, and making their own lives simpler in the process.

And who's installing the fiber? Usually, it's the damn ILECs! They're in competition with themselves. They've been saddled with these billions of buried wires for fifty years, they consider the wireline portion of the business to effectively be a parasite they can't legally get rid of, but they can kill it slowly by forming a new, guerilla business that tempts consumers over to technology that wasn't invented just after the Industrial Fucking Revolution.

So, what this all comes down to is this:

Whether you want voice service or internet (not that I've heard of anyone getting DSL in years,) the copper that's in the ground is not getting any healthier, and if you get a Plain Old Telephone Service line, it might actually be shitty.

I mean, that's not very LIKELY unless you live on a farm, or in a VERY old part of a city, or maybe somewhere out on the east coast that has hurricanes every weekend and is never dry. But if you ARE unlucky, you're stuck with it. If your phone line crackles or pops or hums or cuts out, you're fucked. That's your life, and it's never getting any better. But I can almost promise you that no matter where you are, you can get some kind of broadband that's capable of passing a clean VoIP signal.

Now, didn't I just spend all day talking about how VoIP can't be reliable? Well,

Here's the thing: There is nothing intrinsically unreliable about VoIP, insofar as you can say the same about a jet airliner. Forget to grease a 747's mechanisms on the proper schedule and the control surfaces will jam up during flight and kill you. They are stupefyingly complex and have a million points of failure - but when properly maintained, the damn things just don't die.

I am an expert on VoIP; I will assert this. I have over ten years of experience, not just deploying one company's idiosyncratic product, but getting my hands dirty and my arms elbow-deep into VoIP's asshole, rooting around to discover its truths. I tell you: when configured correctly and given the proper respect, VoIP is rock solid and will just keep working. It's actually very simple, if you don't configure it to be complicated, and it's damn near on par with TDM POTS service, especially if you give it a dedicated network all to itself.

The good forms of POTS-replacement VoIP are like that. If you get Centurylink fiber with included phone service, my impression is that they do a really good job of deploying it. You get a modem that plugs into your fiber and spits out a couple emulated analog phone jacks, but it's battery backed, it's centrally configured by a robust auto-provisioning system, the modem very probably cordons the VoIP traffic off into its own private network segment so it doesn't have to touch the public internet at all, and it's connected to a very reliable backend. Set up like that, it's probably very close to POTS quality.

In fact, many people have been on VoIP FAR longer than they realize. If you've ever had an ADSL "lineshare" - where you have a single line that offers both phone service and DSL - then yeah, it's possible your dialtone was coming from a conventional phone switch... but it's also possible you were just plugged into a VoIP-to-analog adapter.

I've personally been inside of DSLAMs (the gear that produces signals for DSL internet) that had VoIP sections where you programmed in a server address, username and password, and they produced a simulated dialtone just like one of the cheap plastic boxes that my company drops on a customer site, that do the same thing, and which break constantly. A big difference however is that those cost about $75, while the DSLAMs cost $50,000.

More importantly, however, is the fact that those systems are centralized. Like I said in my other post: What makes POTS reliable is that it's all happening in one place. If the complicated bits are all together, then the vendor can maintain and repair them all in one spot, instead of running all over creation to troubleshoot simple problems. If they're centralized, the vendor treats them as capex: equipment they're purchasing, at cost-no-object prices, because it's an investment in themselves. It's not just some box they're dumping on your property that needs to be as cheap and forgettable as possible.

There's really nothing intrinsically wrong with the VoIP concept. It's absolutely capable of delivering high quality service, and in many cases, it can deliver service that's impossible over POTS. I've worked with companies that are out in the boonies, places where they can't even get a conventional phone line, but they can set up an LTE hotspot and connect ten phones through it. Trying to run a business out of ten cell phones is a nightmare scenario, but thanks to VoIP, you can have a really high quality, feature-rich business phone system in the middle of a desert, running off a generator and a SIM card. It... it sucks, but it beats the alternative.

My primary complaints with VoIP are that it uses absurdly old protocols that need to be updated, and that it lacks sufficient self-healing mechanisms. The same can be said about the phone system, however, and if VoIP was treated like the phone system, given the respect it deserved, it would work well.

The problem with the whole POTS-replacement process that's going on right now is just that it's being done as cheaply as possible. Companies are not deploying robust, high quality equipment that's centrally configured and fault-tolerant, and they won't. The reason I just say "voip fucking sucks" is that our nightmarish economic system guarantees that this will only get worse, not better.

Put differently: It is much harder for a business to hurt you with POTS. VoIP, however, is a sword with few safe places to grip.


You must log in to comment.

in reply to @cathoderaydude's post:

we'd actually really love to hear your ideas on centralization at some point, it sounds like you've thought it through more than most people do. we lean the other way but we consider it important to hear counterpoints heh.