I love landline phones (when they work right, which is not the case everythere) because usually they have better sound quality than mobile (which can be good-to-horrible depending on network quality) and almost zero latency.
We often talk of developing nations "leapfrogging" landlines and going straight to mobile, but they're also going straight to fiber. It would be fascinating to see this same type of breakdown for rural fiber infrastructure.
One obvious difference is that fiber can't self-power a phone, but with ubiquitous mobile phones/solar panels/powerbanks that doesn't strike me as a huge show-stopper.
Thanks OP, great post and content. Now I'll go through the world with new eyes!
Obviously excepting the cruelty I love the modern world. I marvel at the organization and patterns.
The Soviet Union was suffering some food crisis in the 90’s. One morning the news was discussing Clinton granting food aid to them. That same morning as my subway crossed by the Dominoe sugar factory there was already a Russian ship docked.
That was such a strange cause and effect moment for me.
It's interesting with East Germany - after German reunification it was covered with fiber. However then DSL became the standard way of going online in Germany, while that doesn't work on fiberand with only few to no consumer priced offerings for FTTH many areas were left out from "fast" Internet for a while, while having the, in theory, superior net.
That's conspiracy. It's more like they could sell SDSL for more money to business customers, as people who need more upload bandwidth gotta pay more, while being able to sell huge numbers in downstream. Plain simple capitalism.
I admit it could be conspiracy but I don't understand why you think the ability to sell a subscription for more money has an effect on the transport medium of the service. They could invent any service they want and run it over fiber.
Do you consider it a conspiracy theory that cable companies sell asymmetric bandwidth? Anytime I see asymmetric bandwidth I see it as an assault against the Hunan right to publish.
That is my reasoning for thinking such conspiratorial thoughts.
The ISP absolutely doesn't care if you publish ... they just don't want somebody to host stuff without paying a premium, since people are willing to pay more for it.
Sounds like a disincentive to publish to me. Remember when all bandwidth was synchronous in the USA. The telecoms were indifferent to which direction the packets went. The cable companies either through technical limitations were the first to offer asynchronous bandwidth as a product to the masses.
Cables companies distribute content and the business model relies on scarcity of access to the content. Their content evolved to leverage centralized advertising model and scarcity of access to content.
This was completely different in the USA during the T1, ISDN, and eventually SDSL.
I always associate the xDSL with the eternal September of the web and it's content.
How one classifies a conspiracy is actually a whole different topic. I would call this more market manipulation.
"Neighbouring Finland offered an analogue telephone exchange as a gift and the Estonian government declined, [...]"
In 1997 even to think about an "upgrade" to an analogue system is a negligence. Estonia already had 3 GSM networks in a country with a population of 1,400,000.
At the time [May 1997] of Ritabell's entering the Estonian market, there were about 64,500 GSM subscribers in the country. 53,000 belonged to EMT and 11,500 to Radiolinja, [...] [0]
EMT (Eesti Mobiiltelefon) claims even bigger numbers:
> Uses in magnesium, aluminium, and electronics manufacturing also hastened atmospheric growth.[10] The 1997 Kyoto Protocol, which came into force in 2005, is supposed to limit emissions of this gas. In a somewhat nebulous way it has been included as part of the carbon emission trading scheme. In some countries this has led to the defunction of entire industries.
The telecom situation in Canada is ridiculous. Even in Greater Vancouver, the 3rd largest metropolitan area in Canada and the 34th in North America, there are dead zones everywhere with major providers. I'm in a suburban development that was mostly built in the last 5-10 years, and there still aren't any towers near here so my bedroom and foyer have no reception (not 1 bar, I mean my phone says "no service, you must be in the wilderness"). I have to stand by a window to make calls. And this is a wood-frame building on flat terrain, next to two major highways.
Telus, Bell and Rogers are ingrained in the minds of Canadians as the only options.
Almost all attempts to change that over the years have failed or have not been able to get sufficent market penetration to become self funding for rapid growth.
Canadian investment funds would rather go and buy U.S. ISPs from failing ventures like Frontier (creating Ziply Fiber) than investing in Canada because they know how bad the situation is. Getting sub-30% market penetration when offering, better, less expensive fiber internet makes the Canadian market uneconomical for deeper investment by insurgent ISPs.
Cell phone coverage in many places is just horrible. You might get good coverage and signal quality, but the network just can't follow when everybody is watching youtube at 4K resolution on their mobile connection. Sometimes even regular phone calls become unreliable (bad sound, interruptions) despite the optimal signal level.
It was a surprise to me, moving from London to Brno in Czechia, to find that both mobile and fixed/home internet was both quicker and much cheaper in this former Communist country. The first house I lived didn't even have a POTS landline installed at all; they skipped that tech being rolled out.
It was a bigger surprise to me, moving to work for a German company in Prague a few years later and thus visiting Germany for work regularly, that German telecomms was much poorer: weaker mobile coverage, slower and much more expensive Internet, and relatively, very few public wifi hotspots.
(I am told that the latter is because of onerous legal restrictions on content that may be downloaded; businesses are liable for anything downloaded by customers on their premises. In comparison, in Czech law it's legal to download copyright material, it's sharing it that's illegal.)
> I love landline phones (when they work right, which is not the case everythere) because usually they have better sound quality than mobile (which can be good-to-horrible depending on network quality) and almost zero latency.
That was the old landlines, of actual POTS, era. Back then, there was a direct electrical circuit literally switched using relays between both participants in a call, with only amplifiers, switches and power sources in between. More modern systems used banks of relays, and then it all went downhill quality-wise.
First, the telco core systems were replaced with digital trunks that had ADCs on both ends of a call in the regional distribution center, and consumers were switched over to ISDN for telephony. Then, the analog/ISDN frontends moved to the curb side where ADSL, VDSL and nowadays G.fast frontends were added to the mix, which were connected to the telco network using fiber. Then, analog and ISDN were shut down, with voice phone calls being migrated to VoIP.
And nowadays, it's the full evolution with GPON - on the telco side, there's only fibers and a single TX/RX pair of transmission modules serving up to 64 customers. What used to require a whole multi-story building can now be done in the space of a better-quality shed. And the analog frontend is only at the CPE, if there is an analog frontend present at all and it's not VoIP softphones or DECT.
My new ISP took this to another level: if you transfer your landline phone number they send you a voice-only SIM card and a 4G router into which you're supposed to plug your landline. Calls to the landline number are redirected to the number in the SIM (with a significant delay). I wonder why they did it this way instead of implementing VoIP, especially since the all-in-one GPON/Wi-Fi router already has a landline port.
In the US, at least in the states I lived in, ISDN telephony was always niche, the price was too high unless you needed it for a remote studio, or you had a teleconference setup or needed the blazing fast 128kbit dial up experience.
Analog phones with a digitizer at the central office or remote terminal is still a very good calling experience with minimal latency. 8-bit u-law @ 8000 Hz isn't great audio quality, but the sampling delay is near zero, and when it was all PRI digital (t1/isdn/etc) switching, multiplexing was done per sample, so there was no significant buffering (a two sample buffer would be sufficient at any switching point).
Mobile uses complex compression with significant sampling delay and sends data in bursts so there's packetization delay. VoIP often uses complex compression (but you can configure for u-law) and is usually 20ms packets, plus you've got to add a jitter buffer to account for packets taking different amounts of time to traverse the network.
Packet switching clearly won over circuit switching, but we've lost the very low latency local calling we used to have, and I don't think anyone is willing to send 1000 packets per second for voice calls to get close to where we were. For long distance calling, probably improved routes that were run for packet switching reduce latency enough to cancel out the increased factors.
I'm not sure what you mean by "better-quality shed" but cabinet sizes today are often less than 1 cubic meter. I guess you know that, but thought it might surprise some people.
One cabinet doesn't really replace a multi-story building. Many cabinets spread out over the service area do the job. But each cabinet can perform all the functions that used to happen in the large building. They just make it economical to perform those functions for 100 - 1000 subscribers, instead of having to centralize them into a single location that serves 5,000 - 100,000.
Those cabinets are not a result of fiber. They were more a necessity of high bandwidth services without fiber. Think ADSL. Fiber actually makes it possible to go back to more centralized service while providing even more bandwidth. You can deploy passive fiber splitters in an old-fashioned pedestal and keep the active optics in the multi-story building for 100,000 subscribers again.
Maybe you meant that the active optics would only fill a small shed's worth of space in that old building. I finally caught up to you.
With 64 XGS-PON customers per SFP+ module, the only active equipment you need to terminate connections for over 60,000 customers fits into a single rack. Large centralized buildings are largely obsolete unless you want to offer services beyond basic Internet access.
The large space savings IMHO come from getting rid of the point to point connections via these massive 2000+ wire cables and their corresponding patch panels and routing infrastructure. With GPON you only need a single SFP module and passive optical splitters in the underground buried junctions, no need for any roadside equipment.
(Worked as a trench digger / roadside equipment box installer once in younger times)
The PSTN has been hybrid fiber/copper for decades. The vast majority of inter-CO traffic is carried on fiber optic lines. Many POTS lines are carried on fiber optic carriers from the central office to a subscriber loop carrier or similar device and then only travel a short distance to the subscriber premises on copper lines.
in the UK it is typically fibre to the box on the street, and then copper from that to the house. providers would love you to subscribe to fibre to the house (so they can charge you exorbitant prices for it), but really YAGNI for most people at least in my experience.
That used to be the case - but it's rapidly becoming standard to be directly connected to fibre. I believe that something like 60% of UK premises have a fibre connection available and I expect that all the providers will push folks off copper as fast as they can.
Because copper is a pain in the arse compared to fibre (for the provider). For the consumer copper has one key positive that fibre just doesn't - it works in a power cut.
In Germany, in particular in the former East, fiber development inadvertently got in the way of fast home Internet access for many years:
Many regions were built out with (at the time of the German reunification) state-of-the-art hybrid fiber/copper systems, with fiber extending even beyond the central office.
Unfortunately, the technology used is incompatible with DSL, the access technology of choice of the former incumbent operator, so any landline deployed using it was ineligible for anything faster than dialup until they either built out a copper link to the central office, or upgraded the existing fiber to GPON (which didn't happen until very recently).
If you think POTS sounds good, you should try ISDN, it'll knock your socks off! Back in the mid-90s I got ISDN BRI (2x64Kbps channel) service in to replace analog modem service. It would connect in well under a second, and while connected at 128K we could receive a phone call (dropping one of the channels, so having voice plus 64K data), and the phone calls sounded quite good.
It used to be that for recording remote interviews or even I believe just general voice recording like audio books, the media companies would have you come into a location that had ISDN, and if you were a real big timer you might have ISDN in your home studio.
That was back in the day when they gave a crap, then the pandemic came along and the media companies seemed to be happy throwing any crappy video chat up for broadcast, not matter how echoy the room you're in or how many drop-outs.
WRT fiber optics, we are just now starting to see Q.com (Century Link) deploying fiber to the neighborhoods. Up in Canada in a similar sized city they deployed fiber back in 2000, but in the states QWest wouldn't do it because, I've been told, it would open up allowing CLECs to put DSL equipment in neighborhoods, and let them cherry-pick neighborhoods to offer service in. Finally a few years ago the city stepped in a ran fiber to every house, which honestly is a better option IMHO.
I've had Quantum Fiber in Seattle for a year or two now. It's fantastic! Symmetrical upload/download, no data cap, 1gbps, and I pay something like $75/mo.
I was paying $100/mo to Comcast for the same thing, though I had 1.2gbps down/35mbps up.
Good to know, I was passingly wondering what the service level was that they were offering. CenturyLink currently can only offer me 60Mbps/5Mbps for $55/mo, until Q.com builds out more of the town.
That compares poorly with the city FTTH which is 1G/1G for $70/mo, 2G/2G for $100, or 10G/10G for $200/mo. A static IPv4 is another $20/mo (which is admittedly pricy for an add-on, but $90/mo for gig with static feels fine to me).
Or if you want something that you can actually get (since it's impossible to get a new ISDN line as a consumer), you could do VoIP with G.722/Opus or cellular with HD Voice (AMR-WB). The quality is honestly on par with ISDN.
And yes, ISDN is still the gold standard for remote voice acting work, but as ISDN lines have gotten more difficult to order a lot of folks have been moving to VoIP. I've heard an app called Source Connect is popular in the broadcast industry, and supposed to provide equivalent quality to ISDN.
ISDN... before it was too expensive, now you can't get it anymore.
VoIP with G.722 is wonderful when it works (you can't just call everybody with it, because both parties must support it). G.711 (A-law or µ-law depending on which continent you are) also works perfectly and gives the same quality of a regular landline. But in all cases the IP transport behind should work well, with very low jitter and not too much latency.
The crowd I ran with in college were the types to order a home-based T-1 service, and it was absurd how unobtainable ISDN was.
We'd receive Sun workstations on our desk (I worked for an ISP) and we'd sort of chuckle at the built-in ISDN interface while we connected the 10BASE2.
We never sold ISDN links; our modems were state-of-the-art 14.4K or you could go Frame Relay, leased line, SMDS.
We did maintain one leased digital 14.4K line, due to some unique circumstance, and I still wish I could forget.
Sure, being a unique situation it also presented unique support challenges, and I seem to recall that the link was the one that most often turned red on our network map, perhaps even teasing us as it flapped up/down for a while, and we sat with bated breath to see whether it really required our undivided attention.
Also I suppose that it was considered the least important of our "dedicated" customer links, being legacy/low-cost, so most of us would've welcomed an upgrade to something more "1990s".
Many providers have ways of passive-aggressively dealing with legacy customers until they upgrade or go to the competition. My DSL ISP sent me into the arms of the cable provider about 12 years ago.
Analog signals won't survive a 1000km undersea cable, but for a few kilometers you don't need amplifiers at all. The loss-per-meter of optical fiber is much lower than a coax cable.
Radio-frequency over fiber (RFoF) signals are usually analog. By intensity-modulating the laser, it's relatively easy to convert a GHz-wide analog signal from electrical to optical and back.
Fiber can indeed be used in a way that is close to coax, or twisted pair, or whatever. It just takes a different [block diagram] of stuff at each end for either configuration.
It's not inherently digital at all.
One of the earliest fiber implementations I ever saw the back end of with my own eyes was a municipal camera system. These were NTSC composite cameras with media converters to convert the electrical signals (over coax) to light signals (over fiber).
It was analog, and dated to 1987 when the building was built. That kind of thing was not new even at that point and did not even begin to push the limits of the mediums.
(And "digital" signals, too: We may tend to think of them as just on/off 1s and 0s, but they're seldom actually that simple in high-speed* signalling systems -- whether electrical or optical.)
*: Whatever it is that "high-speed" means around the time of implementation
I've been enjoying the videos of 'Look Mum No Computer', building his own 60's based, electromechanical telephone exchange, is his Museum of everything else. https://www.youtube.com/watch?v=tK1lH8pjXTo
His second channel has some more dedicated content on telephone systems, https://www.youtube.com/playlist?list=PLKnS0AB2CTN_eu8k8rgaO...
He has spent a fair amount of time building it out, from initially receiving his first exchanging and getting it to work, and adding more to it
It fun just to watch how the system selects lines and routes around the exchange. Other interesting things, like how the dial and engaged tones are generated, and how prerecorded messages are played back to a number that is no longer in service, something like, 'This number no longer exists, please try again'
There's some videos on the AT&T Tech Channel on YouTube that show inside the CO, like this one for the "Speedy cutover service": https://www.youtube.com/watch?v=saRir95iIWk
I really do hope y'all excuse my ignorance, but ...
... what part of the "switchover" process is this? That is to say ...
... I get that you have to cut to switch over but ... what would come next? Does the new system get spliced in? Is it already wired - pending just the cut? (Thus the speed with which it is done) ...
the new system is already spliced in, but is electrically isolated.
the old system can't be isolated like that so they have to physically cut the wires to it before they electrically connect the new system, which is the switch throw at the end of the cutover.
Now we get to a really interesting part. As you drive along rural roads, you'll see what appear to be paint cans, each with a this cable coming out the top.
As you might have guessed, these therefore are not paint cans, but are the enclosures for the loading coils. If there is a 600-pair cable, then you need 600 loading coils (these look something like little spools of thread), and you need to attach each to the feeder cable.
I hadn't seen that config before. (looong-time infra rubbernecker). I assume I've seen variants but am not recalling what they look like.
Yeah - there are colour codes you can use to find the pair that you want to connect, they are arranged in a mystical system of binders. It's like a map.
Back in my telephone days, I used to use this to remember my pair colors.
We Rape Beautiful Young Virgins (for the White Red Black Yellow Violet) and Big Old Gob of Bull Shit (Blue Orange Green Brown Slate). There were other mnemonics other people used.
Back in my younger days when I ran a switching office, I used to rip a few bong hits, put on some music and wire DSX panels on the weekends and just zen out. It was therapeutic.
The way to terminate a 1200-pair cable is the same as eating an elephant: One bite at a time.
Historically (and in the US, at least), telephone wire has been organized into chunks of 25 pairs each, with each pair of wires having a unique color code.
And every 25 pair cable in a given system has the same set of wire colors.
And these 25-pair chunks are easy to terminate on (say) a 66 punch-down block. It takes some time to learn how to get good at it, but not as much time as one might think.
Those 25-pair chunks are organized into 5 different body colors, each with 5 different pair colors.
We're all familiar with the colors of 4-pair cat5 and friends; that's just the first 4 pairs of wire of a 25-pair chunk. (We can tell that they're the first 4 pairs because they all have the color white in common with eachother.)
So, the order for a 25-pair cable is this:
First, the white pairs: Blue, Orange, Green, Brown, Slate.
Then the red pairs: Blue, Orange, Green, Brown, Slate.
Then the black pairs: Blue, Orange, Green, Brown, Slate.
Then the yellow pairs: Blue, Orange, Green, Brown, Slate.
Then the violet pairs: Blue, Orange, Green, Brown, Slate.
Done. A 25 pair cable is terminated.
---
Larger cable are also organized into 25-pair chunks. The colors of the binder strings wrapped around each 25-pair chunk identify it.
This works for cables with up to 600 pairs.
---
1200-pair cable just consists of two 600-pair groups, with each group wrapped in its own colored binder
---
So to begin terminating a 1200-pair cable, first you identify the first group of 600, based on the color of outermost binding wrap. Now the 1200-pair problem is only a 600-pair problem.
Inside of that group of 600, find the first chunk of 25 (the white-blue one). Now your 600-pair problem is only a 25-pair problem, and terminating 25 pairs is easy.
So terminate all 25 pairs, and then move onto the next chunk of 25 (white-orange).
Keep doing this until you've worked through all 24 chunks of 25 pairs in that 600-pair group.
And then just do it again for the other 600 pairs.
As a kid in Los Alamos, the mesa I lived on was far enough from the central office[1] that it had some loading coils between us and them. This meant that we couldn't get DSL on that mesa. The mesa across the way didn't have the loading coils, and thus was able to get DSL. I remember the jealousy of my friend being able to download DOOM WADs in a couple of seconds, while I had to schedule them, or convince my dad to download them on the much faster connection at LANL.
Eventually we managed to get either an ISDN line or a T1 line, I cannot remember which, but shortly thereafter Cable internet became available, and rendered the whole problem moot.
I didn't know that these cables are all pressurized with air to keep leaks out (and to dry out any leaks that do happen). To me, that was honestly the most interesting thing in the article.
87 comments
[ 5.1 ms ] story [ 187 ms ] threadBut it's time to switch to fiber optics!
PS. But, can you do fence-wire comms with fiber? :)
Maybe with a transparent plastic fence :-D
(But your fence bend radius is limited, though ... :)
One obvious difference is that fiber can't self-power a phone, but with ubiquitous mobile phones/solar panels/powerbanks that doesn't strike me as a huge show-stopper.
Thanks OP, great post and content. Now I'll go through the world with new eyes!
EDIT: found the fiber page http://cityinfrastructure.com/single.php?t=Fibre%20Optic%20C...
I'd be interested to know - qua infra - what their power grids are looking like ...
https://theconversation.com/estonia-is-a-digital-republic-wh...
PS. Also, interesting that there - apparently - is a "second hand" market for used phone systems. As in, wholesale.-
The Soviet Union was suffering some food crisis in the 90’s. One morning the news was discussing Clinton granting food aid to them. That same morning as my subway crossed by the Dominoe sugar factory there was already a Russian ship docked.
That was such a strange cause and effect moment for me.
Do you consider it a conspiracy theory that cable companies sell asymmetric bandwidth? Anytime I see asymmetric bandwidth I see it as an assault against the Hunan right to publish.
That is my reasoning for thinking such conspiratorial thoughts.
Cables companies distribute content and the business model relies on scarcity of access to the content. Their content evolved to leverage centralized advertising model and scarcity of access to content.
This was completely different in the USA during the T1, ISDN, and eventually SDSL.
I always associate the xDSL with the eternal September of the web and it's content.
How one classifies a conspiracy is actually a whole different topic. I would call this more market manipulation.
In 1997 even to think about an "upgrade" to an analogue system is a negligence. Estonia already had 3 GSM networks in a country with a population of 1,400,000.
At the time [May 1997] of Ritabell's entering the Estonian market, there were about 64,500 GSM subscribers in the country. 53,000 belonged to EMT and 11,500 to Radiolinja, [...] [0]
EMT (Eesti Mobiiltelefon) claims even bigger numbers:
By 1996, EMT had 100,000 subscribers. [1]
[0] https://books.google.lt/books?id=SrFXo6qDRjcC&pg=PA152
[1] https://www.telia.ee/en/uudised/celebrating-25-years-of-mobi...
(Sorry, kinda replying to the below as well)
Great for the SF6 to supercritical CO2 transition (which hitachi is already helping china with) https://www.technologyreview.com/2024/09/02/1103398/greenhou...
> Uses in magnesium, aluminium, and electronics manufacturing also hastened atmospheric growth.[10] The 1997 Kyoto Protocol, which came into force in 2005, is supposed to limit emissions of this gas. In a somewhat nebulous way it has been included as part of the carbon emission trading scheme. In some countries this has led to the defunction of entire industries.
https://en.wikipedia.org/wiki/Sulfur_hexafluoride
I could go on about what the long term solution is but who wants to hear rants from a kiddo ;)
(1) https://www.cbc.ca/news/canada/prince-edward-island/pei-cell...
Almost all attempts to change that over the years have failed or have not been able to get sufficent market penetration to become self funding for rapid growth.
Canadian investment funds would rather go and buy U.S. ISPs from failing ventures like Frontier (creating Ziply Fiber) than investing in Canada because they know how bad the situation is. Getting sub-30% market penetration when offering, better, less expensive fiber internet makes the Canadian market uneconomical for deeper investment by insurgent ISPs.
The Canadian consumer is easily swayed by Bell/Telus/Rogers propoganda like https://web.archive.org/web/20140402040045/http://fairforcan...
911 is an entirely different class of service.
It was a bigger surprise to me, moving to work for a German company in Prague a few years later and thus visiting Germany for work regularly, that German telecomms was much poorer: weaker mobile coverage, slower and much more expensive Internet, and relatively, very few public wifi hotspots.
(I am told that the latter is because of onerous legal restrictions on content that may be downloaded; businesses are liable for anything downloaded by customers on their premises. In comparison, in Czech law it's legal to download copyright material, it's sharing it that's illegal.)
That was the old landlines, of actual POTS, era. Back then, there was a direct electrical circuit literally switched using relays between both participants in a call, with only amplifiers, switches and power sources in between. More modern systems used banks of relays, and then it all went downhill quality-wise.
First, the telco core systems were replaced with digital trunks that had ADCs on both ends of a call in the regional distribution center, and consumers were switched over to ISDN for telephony. Then, the analog/ISDN frontends moved to the curb side where ADSL, VDSL and nowadays G.fast frontends were added to the mix, which were connected to the telco network using fiber. Then, analog and ISDN were shut down, with voice phone calls being migrated to VoIP.
And nowadays, it's the full evolution with GPON - on the telco side, there's only fibers and a single TX/RX pair of transmission modules serving up to 64 customers. What used to require a whole multi-story building can now be done in the space of a better-quality shed. And the analog frontend is only at the CPE, if there is an analog frontend present at all and it's not VoIP softphones or DECT.
Analog phones with a digitizer at the central office or remote terminal is still a very good calling experience with minimal latency. 8-bit u-law @ 8000 Hz isn't great audio quality, but the sampling delay is near zero, and when it was all PRI digital (t1/isdn/etc) switching, multiplexing was done per sample, so there was no significant buffering (a two sample buffer would be sufficient at any switching point).
Mobile uses complex compression with significant sampling delay and sends data in bursts so there's packetization delay. VoIP often uses complex compression (but you can configure for u-law) and is usually 20ms packets, plus you've got to add a jitter buffer to account for packets taking different amounts of time to traverse the network.
Packet switching clearly won over circuit switching, but we've lost the very low latency local calling we used to have, and I don't think anyone is willing to send 1000 packets per second for voice calls to get close to where we were. For long distance calling, probably improved routes that were run for packet switching reduce latency enough to cancel out the increased factors.
One cabinet doesn't really replace a multi-story building. Many cabinets spread out over the service area do the job. But each cabinet can perform all the functions that used to happen in the large building. They just make it economical to perform those functions for 100 - 1000 subscribers, instead of having to centralize them into a single location that serves 5,000 - 100,000.
Those cabinets are not a result of fiber. They were more a necessity of high bandwidth services without fiber. Think ADSL. Fiber actually makes it possible to go back to more centralized service while providing even more bandwidth. You can deploy passive fiber splitters in an old-fashioned pedestal and keep the active optics in the multi-story building for 100,000 subscribers again.
Maybe you meant that the active optics would only fill a small shed's worth of space in that old building. I finally caught up to you.
(Worked as a trench digger / roadside equipment box installer once in younger times)
The PSTN has been hybrid fiber/copper for decades. The vast majority of inter-CO traffic is carried on fiber optic lines. Many POTS lines are carried on fiber optic carriers from the central office to a subscriber loop carrier or similar device and then only travel a short distance to the subscriber premises on copper lines.
Because copper is a pain in the arse compared to fibre (for the provider). For the consumer copper has one key positive that fibre just doesn't - it works in a power cut.
Many regions were built out with (at the time of the German reunification) state-of-the-art hybrid fiber/copper systems, with fiber extending even beyond the central office.
Unfortunately, the technology used is incompatible with DSL, the access technology of choice of the former incumbent operator, so any landline deployed using it was ineligible for anything faster than dialup until they either built out a copper link to the central office, or upgraded the existing fiber to GPON (which didn't happen until very recently).
Definitely, I was mainly thinking about those thousands of lines coming out from the CO...
It used to be that for recording remote interviews or even I believe just general voice recording like audio books, the media companies would have you come into a location that had ISDN, and if you were a real big timer you might have ISDN in your home studio.
That was back in the day when they gave a crap, then the pandemic came along and the media companies seemed to be happy throwing any crappy video chat up for broadcast, not matter how echoy the room you're in or how many drop-outs.
WRT fiber optics, we are just now starting to see Q.com (Century Link) deploying fiber to the neighborhoods. Up in Canada in a similar sized city they deployed fiber back in 2000, but in the states QWest wouldn't do it because, I've been told, it would open up allowing CLECs to put DSL equipment in neighborhoods, and let them cherry-pick neighborhoods to offer service in. Finally a few years ago the city stepped in a ran fiber to every house, which honestly is a better option IMHO.
I was paying $100/mo to Comcast for the same thing, though I had 1.2gbps down/35mbps up.
That compares poorly with the city FTTH which is 1G/1G for $70/mo, 2G/2G for $100, or 10G/10G for $200/mo. A static IPv4 is another $20/mo (which is admittedly pricy for an add-on, but $90/mo for gig with static feels fine to me).
And yes, ISDN is still the gold standard for remote voice acting work, but as ISDN lines have gotten more difficult to order a lot of folks have been moving to VoIP. I've heard an app called Source Connect is popular in the broadcast industry, and supposed to provide equivalent quality to ISDN.
VoIP with G.722 is wonderful when it works (you can't just call everybody with it, because both parties must support it). G.711 (A-law or µ-law depending on which continent you are) also works perfectly and gives the same quality of a regular landline. But in all cases the IP transport behind should work well, with very low jitter and not too much latency.
We'd receive Sun workstations on our desk (I worked for an ISP) and we'd sort of chuckle at the built-in ISDN interface while we connected the 10BASE2.
We never sold ISDN links; our modems were state-of-the-art 14.4K or you could go Frame Relay, leased line, SMDS.
We did maintain one leased digital 14.4K line, due to some unique circumstance, and I still wish I could forget.
May I ask why?
Also I suppose that it was considered the least important of our "dedicated" customer links, being legacy/low-cost, so most of us would've welcomed an upgrade to something more "1990s".
Many providers have ways of passive-aggressively dealing with legacy customers until they upgrade or go to the competition. My DSL ISP sent me into the arms of the cable provider about 12 years ago.
https://en.wikipedia.org/wiki/Radio_over_fiber
https://www.rp-photonics.com/radio_and_microwave_over_fiber....
It's not inherently digital at all.
One of the earliest fiber implementations I ever saw the back end of with my own eyes was a municipal camera system. These were NTSC composite cameras with media converters to convert the electrical signals (over coax) to light signals (over fiber).
It was analog, and dated to 1987 when the building was built. That kind of thing was not new even at that point and did not even begin to push the limits of the mediums.
(And "digital" signals, too: We may tend to think of them as just on/off 1s and 0s, but they're seldom actually that simple in high-speed* signalling systems -- whether electrical or optical.)
*: Whatever it is that "high-speed" means around the time of implementation
His second channel has some more dedicated content on telephone systems, https://www.youtube.com/playlist?list=PLKnS0AB2CTN_eu8k8rgaO... He has spent a fair amount of time building it out, from initially receiving his first exchanging and getting it to work, and adding more to it
It fun just to watch how the system selects lines and routes around the exchange. Other interesting things, like how the dial and engaged tones are generated, and how prerecorded messages are played back to a number that is no longer in service, something like, 'This number no longer exists, please try again'
The channel is about the fully functional Central Office turned Museum in Seattle
There's a fully functional system deployed in Seattle as a Museum
... what part of the "switchover" process is this? That is to say ...
... I get that you have to cut to switch over but ... what would come next? Does the new system get spliced in? Is it already wired - pending just the cut? (Thus the speed with which it is done) ...
the old system can't be isolated like that so they have to physically cut the wires to it before they electrically connect the new system, which is the switch throw at the end of the cutover.
As you might have guessed, these therefore are not paint cans, but are the enclosures for the loading coils. If there is a 600-pair cable, then you need 600 loading coils (these look something like little spools of thread), and you need to attach each to the feeder cable.
I hadn't seen that config before. (looong-time infra rubbernecker). I assume I've seen variants but am not recalling what they look like.
Terminating that must be a lot of fun. Please tell me they have some clever device that makes this task easier.
The model train club I used to be a member of was full of old telephone guys, and they'd share old stories
We Rape Beautiful Young Virgins (for the White Red Black Yellow Violet) and Big Old Gob of Bull Shit (Blue Orange Green Brown Slate). There were other mnemonics other people used.
https://en.wikipedia.org/wiki/25-pair_color_code
Back in my younger days when I ran a switching office, I used to rip a few bong hits, put on some music and wire DSX panels on the weekends and just zen out. It was therapeutic.
Historically (and in the US, at least), telephone wire has been organized into chunks of 25 pairs each, with each pair of wires having a unique color code.
And every 25 pair cable in a given system has the same set of wire colors.
And these 25-pair chunks are easy to terminate on (say) a 66 punch-down block. It takes some time to learn how to get good at it, but not as much time as one might think.
Those 25-pair chunks are organized into 5 different body colors, each with 5 different pair colors.
We're all familiar with the colors of 4-pair cat5 and friends; that's just the first 4 pairs of wire of a 25-pair chunk. (We can tell that they're the first 4 pairs because they all have the color white in common with eachother.)
So, the order for a 25-pair cable is this:
Done. A 25 pair cable is terminated.---
Larger cable are also organized into 25-pair chunks. The colors of the binder strings wrapped around each 25-pair chunk identify it.
This works for cables with up to 600 pairs.
---
1200-pair cable just consists of two 600-pair groups, with each group wrapped in its own colored binder
---
So to begin terminating a 1200-pair cable, first you identify the first group of 600, based on the color of outermost binding wrap. Now the 1200-pair problem is only a 600-pair problem.
Inside of that group of 600, find the first chunk of 25 (the white-blue one). Now your 600-pair problem is only a 25-pair problem, and terminating 25 pairs is easy.
So terminate all 25 pairs, and then move onto the next chunk of 25 (white-orange).
Keep doing this until you've worked through all 24 chunks of 25 pairs in that 600-pair group.
And then just do it again for the other 600 pairs.
[]: http://cityinfrastructure.com/OutsidePlant/Webfiles/colorcod...
Eventually we managed to get either an ISDN line or a T1 line, I cannot remember which, but shortly thereafter Cable internet became available, and rendered the whole problem moot.
[1] Los Alamos had one central office, with a big microwave relay in the top of it: https://i.imgur.com/Lcfj9oA.jpeg