[HN Gopher] ISDN: The Untravelled Path of Mobile Computing (2015)
       ___________________________________________________________________
        
       ISDN: The Untravelled Path of Mobile Computing (2015)
        
       Author : todsacerdoti
       Score  : 25 points
       Date   : 2021-05-05 16:03 UTC (6 hours ago)
        
 (HTM) web link (blog.gatunka.com)
 (TXT) w3m dump (blog.gatunka.com)
        
       | g051051 wrote:
       | I had a bonded ISDN setup with my company paying half of the
       | fees, back in 1997 or so. I got to work from home, over a rock-
       | steady 128kbit connection, on my home Linux PC connected to my HP
       | workstation in the office. It was a magical time.
        
       | hlandau wrote:
       | I wonder, were there any countries where ISDN BRI became a
       | "default" service, i.e. where it was normal for it to be
       | installed in every household to the same extent POTS was?
        
         | ahofmann wrote:
         | ISDN was used a lot in Germany and was disabled just a few
         | years ago. I still remember the agony of getting my Linux PC to
         | work with the Fritz!Card (on the ISA bus) to finally be able to
         | be online for 3 cents a minute on the Internet. ISDN was really
         | fast and extremely stable though.
        
           | LargoLasskhyfv wrote:
           | Hehehe. Wrong move :-) Should have used something with either
           | a chip from Cologne Chip Design, or the other big asian
           | player whose name I can't remember at the moment. Anyways,
           | both were way more affordable and had stable drivers.
           | Basically looking at what Isdn4Linux, and later HISAX
           | supported well.
        
         | the_only_law wrote:
         | One thing that still confuses me about ISDN is that despite
         | most of it being standardized in a series of ITU
         | specifications, I see a lot of phones and other TE's advertise
         | support for vendor specific protocols and other standards
         | (National ISDN protocols, special 5ESS protocols and later
         | things like DSS1)
         | 
         | What I don't get is the signalling is all standardized (LAPD,
         | X.25 or maybe Frame Relay at layer 2, and Q.931 at layer 3) and
         | the actual data should just transparent to the protocol and
         | only relevant to the devie function (I imagine a typical voice
         | call would just alaw or ulaw data). So unless there's something
         | going in the upper layers of an established call, why the need
         | for various proprietary protocols that from the sounds of it
         | are incompatible.
         | 
         | There are devices and tools to simulate this proprietary
         | protocols, so I suppose in theory I could reverse them, but
         | they're either rare to come by or very expensive for a random
         | hobbyist such as myself.
        
         | ac29 wrote:
         | I seem to recall in the US, there was a narrow window of time
         | where ISDN was interesting - one of my parents had an ISDN
         | connection to their office installed in our house. Within a
         | couple years, the vastly faster ADSL became available followed
         | shortly thereafter by cable internet.
        
           | toast0 wrote:
           | IIRC, in California, ISDN only had a per minute tarrif (which
           | would run at 2x if you used both B channels) and you usually
           | had to pay your ISP more too, although I heard after 56k
           | modems were widely deployed, you could use one B channel and
           | usually get a 64k (both ways) connection for regular dialup
           | cost.
        
         | pantalaimon wrote:
         | ISDN was pretty prevalent in Germany, but it always would cost
         | a premium compared to POTS
        
           | LargoLasskhyfv wrote:
           | Not really true from the times I remember, because heavily
           | subsidised.
        
       | tibbydudeza wrote:
       | Before we got ADSL we had ISDN when our local carrier introduced
       | unlimited weekend calling only billing for the first minute as
       | long as you never dropped the call.
       | 
       | I used channel bonding so I could get 128K using a Linux server
       | as the dialer/proxy as the Windows drivers for the ISDN card was
       | just too darn flaky.
       | 
       | Good times.
        
       | the_only_law wrote:
       | Heh, I've been doing a lot of reading on ISDN lately, working on
       | a user mode ISDN stack myself. Much more than I should tbh. I
       | wonder if the ones pictured in the article use BRI or PRI
       | interfaces, I'd have to guess PRI personally.
       | 
       | I actually have a phone like that on my eBay watchlist, except
       | it's a video phone (using H.320 I imagine) I've been meaning to
       | pick it up, I've have a couple unexpected expenses recently.
        
         | icedchai wrote:
         | PRI was similar to a T1. I kinda doubt they had that at the
         | payphone level.
        
           | the_only_law wrote:
           | Perhaps you're right. I was thinking that if this was part of
           | a network of pay phones that they would want to be able to
           | handle more than just the 2 calls per line an BRI would allow
           | with less infrastructure (though I think lower rate calls are
           | possible through additional multiplexing)
        
         | hlandau wrote:
         | I'd guess it used BRI. PRI would probably have been considered
         | too expensive to justify running to a payphone. My guess is the
         | BRI socket on that payphone is only operable when you're not
         | making a call, so the 2x64k BRI line can be assigned to the
         | phone or passed through to that socket.
         | 
         | You're working on a user-mode ISDN stack - this sounds
         | extremely interesting! Care to drop me a line?
         | https://www.devever.net/~hl/contact
        
       | _joel wrote:
       | I recall my dad getting bonded ISDN at his shop years ago
       | (1996/7?) the jump from 28.8kbit (x2 56 was just out but only
       | supported on some ISPs iirc) to 128kbit seemed like another
       | world. So much so I'd lug my own PC down and pretty much spend
       | all my Saturdays there. Good times
        
         | cotillion wrote:
         | I got very lucky about then when my father got his work to pay
         | for bonded ISDN to our house in the middle of nowhere. Being
         | able to connect to BBSes at 128kbit was insane.
         | 
         | But the big thing with ISDN and the internet was the latency. A
         | bunch of mudding friends brought over their computers all the
         | time since they could suddenly play with a latency almost as
         | good as the older kids who played from their university dorms.
        
         | aequitas wrote:
         | The 2 x 64kbit was nice and all, but being able to surf all
         | evening (reduced tariff) on 1 line and have the house phone
         | available for incoming calls (for my parents) at the same time,
         | that was the game changer for me. No more, "just 30 minutes
         | internet tonight kids, I'm expecting an important call", just
         | dial in at 8 and start the download manager for the entire
         | evening.
        
         | vardump wrote:
         | 128 kbps ISDN had just about 40 ms latency. A significant
         | improvement to a modem, which had about 100 ms+ IIRC, often a
         | lot more, even 500 ms+.
         | 
         | That low latency made ISDN feel fast, sometimes even better
         | than the first ADSL 512-1024 kbps modems that used interleave
         | error correction method. Interleaving caused latency to be
         | about 60-70 ms IIRC, making the higher bandwidth ADSL
         | connection feel slower.
         | 
         | Of course then there was 144 kbps IDSL, which was really 2x 64
         | kbps ISDN + 16 kbps ISDN control channel.
        
           | jaywalk wrote:
           | IDSL was the first "high speed" connection I ever had at
           | home. What a dream it was compared to dialup! Always on and
           | 3-4x faster than the best dialup connection could do.
        
           | toast0 wrote:
           | It feels like you're adding about 30-50ms onto the base
           | latencoes here. Are you/were you fairly far from any major
           | internet exchanges when ISDN and ADSL arrived?
        
             | vardump wrote:
             | IIRC, those were ping latencies to the first hop discovered
             | by using traceroute.
             | 
             | ADSL was about 30-35 ms with fast path in use instead of
             | interleave.
        
               | toast0 wrote:
               | My experience with fast path ADSL was about 8 ms. I don't
               | think I was on interleaved ADSL, but interleaved VDSL2 is
               | about 20ms.
        
               | vardump wrote:
               | Was this in year 2000 or so?
        
               | toast0 wrote:
               | I had ADSL1 at 6m/768k fast mode roughly 8ms latency from
               | 2004-201x; the latency was pretty much constant (it did
               | get slightly better when I replaced my modem after the
               | first one died, but the improvement was less than a ms,
               | but measurable). ADSL1 didn't change much after release
               | afaik, changes came through ADSL2 and VDSL and such. The
               | scourge that is PPP over Ethernet did become more popular
               | later in the life of DSL though.
        
           | wmf wrote:
           | I remember my ISDN having closer to zero latency. It was kind
           | of amazing watching Web pages jump onto the screen.
        
       | LargoLasskhyfv wrote:
       | Onyl possible in Japan. Because less vandalism (regarding the
       | example public payphone pictured there, whith RJ-11 and RJ-45
       | sockets)
        
       ___________________________________________________________________
       (page generated 2021-05-05 23:02 UTC)