Slashdot Mirror
Ask Slashdot: What Is the Future of Old Copper Pair Technology?
p00kiethebear writes "My father works for a large corporation that licenses ISDN lines (among a plethora of other services) including T1 and T3 technology. Surprisingly there are still large companies that use fifty year old T1 technology to handle their voice and data use. My father's 30 year career has been almost exclusively in helpdesk / troubleshooting T1 / ISDN technology and both he and I are worried about the future. Cable modems and DSL have replaced ISDN in most cases and it's now an archaic solution reserved for voice actors, tech support-terminal workers, large companies that need voice and video conferencing, and data and private users too far from the loop for DSL or Cable. My dad is still 15 years from retirement. Is twisted copper going the way of the dodo or is it here to stay for the foreseeable future?"
For low-latency and lossless point-to-point across town, we couldn't find ANY ISP's connection technology that could beat the T1.
Expensive, but rock solid and quick (vs fast).
THL phish sticks
Same holds true in Canada. Though they have been rolling out FTTN for the last few years, but in the end last mile is still copper.
Om, nomnomnom...
You seem to be focused on BRI ISDN which is what is used by those you referenced (TV remotes, voice actors, etc). It is an extremely low bandwidth connection (128 Kbps) but "it works" and is probably not going away anytime soon. PRI is probably much more prevalent. PRI is what I would consider the T1 of ISDN. It is commonly used for enterprise PBX systems, and I definitely don't see it going away anytime soon. The only other realistic option I see at present is SIP, but even then unless it's delivered over fiber SIP services are still probably going to come in over some kind of copper medium (be it T1, etc). Some companies are moving to fiber, but there is usually considerably more cost associated with bringing fiber to the premises as compared to copper which likely is already on premise.
My company has fiber on premise for IP, but we still have PRIs from the LEC for our voice service. Any time you bring voice in over an IP transport (as in SIP), you have to make sure the IP network has proper QoS, etc whereas PRI "just works". PRI is usually more expensive, but not overly so. When we replaced our PBX a few years ago we considered SIP, but when we presented the various options to the powers that be, they chose to stick with PRI because it has a proven track record whereas SIP was just gaining traction in the market.
I think in 15 years you will definitely see fiber steal a large market share of those customers that are currently using copper, but I think there will still be plenty of copper around.
Like someone else commented, the poster uses terms "Copper" and "ISDN" interchangeably. However, with the inclusion of terms like T1/T3, it's clearly about "what can an old telco-guy do in this newfangled IP-based world with 15 years before retirement". Copper here is a misnomer, a lot of stuff can happen over copper (DSLs being the most obvious example).
I have some familiarity in just how dead the technology is. We have a big customer who just placed a big order for Cisco's PVDM digital modems. Why "big", if the tech is dying? Well, that stuff is going to end-of-sale after this summer and they have lot of legacy systems around the globe that dial in (machine-to-machine stuff, and not easily upgradeable everywhere at once). They are moving to IP-based systems but cannot really do that fast enough. Anyway, one of the biggest vendors of network equipment just decided that they aren't going to sell modems that can talk directly to E1/T1 line (analog 2-port models are still in the selection though). I don't know that anyone else is selling such stuff either (Alcatel maybe?). That technology had it's day, but it's long gone.
There might of course be places where, due to signaling constraints, you need to run a E1/T1, but it doesn't really use any of the features. You just run PPP over that link and be done with it - no one cares about the intricasies of Q.931 framing or setting up calls for such links. Even in telephony, it will continue to have some uses, for example many PBX systems still only provide E1/T1 uplink - even if it's going to be used just to connect couple of feet to the SIP gateway right at the next rack.
Frankly, your father has two choices: Either
a) Get entrenched into some niche that really can keep on going with ISDN-based technologies for the next 15 years - you know, maintain job security by being the "only one left who understands this piece of legacy junk that we cannot migrate away from fast". Frankly, I find such positions hard to imagine - sure, maybe if he was retiring in this decade, it could work, but hardly in the 2020's.
or
b) Join the IP world. Frankly, I would think that with a reasonable effort he could still become an expert in VoIP - you still need skills like provisioning (for QoS), codecs (even the G.711a/mu-law is relevant), and so on. Lot of the concepts in SIP are still based on the good old stuff from telco days. You just need to wrap your head around the concept that instead of TDM sending each frame at exactly right intervals, you get packets that might occasionally get lost or routed wrongly or arrive out-of-order...And frankly, you also don't need to care anymore about stuff like SPID's or TEIs. Which I would think of a relief.
We use FIOS for our internet connectivity, but we still rely on MPLS over T1s to interconnect our offices and handle our VOIP traffic. VPN over the public internet simply had too much latency to be useful. It's archaic, but it works.
Which is utterly incapable of replacing copper pair for the simple reason known as "physics".
Really? There are millions of people running their successful small businesses on various combinations of GSM, CDMA, 3G, 4G, 802.11, and WiMax.
Most small firms don't have huge bandwidth or speed needs. They need a payment terminal to process credit cards, an email client to handle customer service, a web browser to look up parts and order materials online, someone to advertise on Google and Facebook and Twitter, and maybe an online QuickBooks account. But most of these people spend the bulk of their days on the sales floor or in the workshop. They don't need a land line, or a server rack, or even a T1. These small businesses used to use dial-up, but everything they need now is available via cellular towers and iPhones.
They don't need the constant reliability that a remote technology worker might need. If the network is down, there's always another order to be filled, a floor to be swept, or shelves to be straightened.
Their buildings still have Cat-3 wiring and old POTS phones gathering dust in the backrooms. But there's no dialtone on them anymore, and there won't be.
John
There's already terminology for this, though not 'single word'
- long = high latency
- fat = high throughput
So a satellite connection would be a 'long thin pipe' usually, while a VDSL2 or fibre connection would be 'short fat'.
See http://en.wikipedia.org/wiki/Bandwidth-delay_product for background - as you probably know, the product of bandwidth (throughput) and delay is also the amount of data buffered in the connection.