We are going to install a backup data line in an off-site location in case of disaster at our main office. The bandwidth can be relatively low, as it would probably be connecting 10 people at the most to the internet.
Is there any significant difference in the level of service/resiliency of the network in general between dsl and cable? If a medium to major disaster were to hit the area, would one infrastructure be better able to handle the increased traffic/possible damage than the other? An thoughts/opinions are appreciated.
Cable in my area does not have battery backup. A power outage anywhere between you and the main site means your connection is toast.
ADSL has battery backup and central office generator backup. ADSL service thru a local my telco provider includes dialup service as well so if your broadband connection went down you could still dial into the network with your v.92 modem.
Trace the wires yourself. Cable center is often nothing more than a metal shack located on some back lot surrounded with a chain link fence. No power backup. No security. No apparent (serious) surge protection. Furthermore, the cable is dependent on amplifiers on utility poles - also no security and no power backup. Cable only provides what is is minimally required to provide service and is completely dependent on every part of the electric company system working. Intermediate amplifiers in selective sections of town without power can take out your cable service.
The phone company, as others have noted, are in hardened brick buildings (some were hardened for war), with battery backup (for four hours), and with generator backup (for days of power). Furthermore, DSL and ISDN is already being obsoleted with a dedicated fiber optic from your building to theirs (no intermediate amplifiers), terminates inside the same hard building, and with days of backup power.
Of course both solutions (cable or telco) require you have on site power and support that is just as reliable.
Actually, neither makes a *good* disaster recovery service.
Cable has the power is the distribution system issues (widespread power outages take the whole cable system down).
DSL if fed directly from the Central Office has an advantage there. C.O.'s are battery backed up with standby generators with many hours of diesel fuel in the tanks. They can run indefinitely as long as fuel can be delivered every few days.
DSL fed from Remote Terminals (R.T.'s), while battery backed up doesn't have standby generator power available. I've seen a few R.T.'s "flatline" after the batteries died after an extended power outage of just a few hours. These things are supposed to be designed to run for 8 hours or more but many have either been poorly maintained or have too much equipment installed and have much shorter runtimes. Sometimes as short as just two or three hours. There are facilities for powering these terminals from portable generators but they need to be dispatched and that takes time, sometimes more time than the batteries will last.
But Cable and DSL not high priority services. That is, they aren't high on the list of things to be restored when there is a crisis. Cable companies are still in the "it's only TV, it'll be back on when we get around to it" mode of operation. Unfortunately, the Phone Companies are starting to slip into that mode as well.
One person suggested ISDN. That's a possibility but ISDN still isn't a high priority restoral item either.
FWIW; I can testify that during the Big Blackout, a couple years ago, my home DSL connection stayed the whole time, more than 24 hours. I had my Linksys and DSL modem plugged into a decent UPS that serves my big desktop system. When the lights went out I turned off the desktop machine and fired up my laptop. That gave me 3 hours of use, rationed over several hours. When that battery died I jacked into the UPS and it carried me thru the rest of the outage.
Depending on how critical your business is, your contingecy plan needs a contingency plan. sh*t happens, and it's completely outside your control. been there, done that.
had the best article I've seen in a long while, it was an update based on lessons learned in Manhattan on 9-11.
I'm in NYC and was one of the later areas to come back. My DSL modem and linksys pox were/are on UPS with my desktop system. When the lights went out I shut down my big system and preceeded to use a laptop, with the USP to use the internet. never dropped a beat.
Most (all?) cable networks were designed for entertainment. So they omit things such as redundant circuits and emergency power/battery backup. The telcos tend to be a lot more anal (good thing) about network quality and reliablility.
I imagine that there is still a public saftey requirement in the FCC and state PSC regulations, at least for voice calls. I have no idea what priority they require for the digitial side of the network..
Speaking of which, there was a massive US North East powerfail a few years ago, including MAE East. How did the Internet fare? particularly Europe/NAm traffic that would be mostly routed through submarine optical cables from the Boston area.
I live outside the balckout zone, and don't recall any difficulties.
The Internet is designed to route around missing components. It is not a single-path heirarchy except near the fringes. Even modest-sized ISPs are multihomed.
Never use (post 1970) AT&T as an example of quality or reliability. Once the telcos broke up, AT&T took the MBAs; phone people went with the baby Bells. Numerous regional and nationwide blackouts by AT&T throughout the 1980s demonstrates how bad AT&T became. Look at their top management. Business school graduates without product experience. No experience working where the work gets done. Therefore AT&T preached a lie - the purpose of a business is it profits - the customer be damned. That mentality explains why AT&T even blamed their own unions for blackouts created by technically naive AT&T management.
AT&T is an benchmark example of how to run an unreliable business. Their top management does not come from where the work gets done and blames everyone (except top management) for their failures. Best example of this was Robert Allen. Classic example of a business based upon principles taught in business schools - where the product is only a messy necessity.
BTW, where does > This was true when there was just The Phone Company (AT&T) [or at
Here in DC last summer we lost electricity for nearly a week. The DSL line as well as the POTS line it shared stayed live the entire time. That and the telco repairs to lines of neighbors were repaired much more quickly than cable or even electric. YMMV of course.
There are any number of additional variables to consider. Mainly just how wide a disaster are you talking about? If it's just your building going up in flames that's one thing. But if it's an outage spanning an entire region there are a whole range of other issues to consider. As in, if all your customers are local it won't do you much good to be up and running if they're still out of service.
As for which technology to use I'd go with telco-based DSL. The infrastructure for telco is generally geared toward much more reliability and maintenance than that of broadband cable. If repairs are being made for phone service it's likely any telco-based data services will benefit at the same time. Cable, on the other hand, doesn't generally have the same degree of commitment to service or anywhere near the same level of personnel geared toward maintenance.
Do baby Bells only install enough fiber optics to meet the load - no backup spares? When one fiber optic in Newark is accidentally broken, then the entire NYC long distance fails - no spare capacity. Is baby Bell management so ignorant of how the power backup systems work as to switch over to batteries? Four hours later, even the Concord must land at another airport with a runway slightly too short because, again, all NYC AT&T service is lost. Would the baby Bells then blame unions before learning why the failure happened? Do the baby Bells install upgrade software that was never tested so that the entire nation's long distance service cascades into a crash? Would they so mismanage their satellite operations as to let their birds be damaged by a solar event? And when these failures happen, does the corporate president (Robert Allen) quickly go to the press blaming the unions?
Did the baby Bells sell more cell phones knowing full well that they did not have sufficient cell phone towers in NYC to support all those phones? Go from one cell to another, get cutoff because that other cell has too many users. AND not be able to call back that so important client. This also was AT&T.
Did the baby Bells buy two cable companies? Then learn, after the fact, that all wiring in those cable networks was electrically insufficient. The entire cable network had to be completely replaced. AT&T did not have the money. So did the baby Bells sell off those cable companies for hundreds of millions of dollars less than they originally paid? Notice the numbers. Notice failures repeatedly traceable to AT&T business school trained management. Failures in AT&T that the baby Bells do not create.
There exists a night and day difference between the baby Bells and AT&T. AT&T philosophy has always been based upon the business school principles (because AT&T is dominated by people with MBA degrees). The Economist Magazine once stated no other company in the history of free market economics destroyed shareholder value more than what AT&T did to NCR. Again more facts based upon numbers. When it comes to bad - really bad - companies, then a benchmark is AT&T. Literally every business that AT&T touches is a long term loser. AT&T routinely stifles innovation. One need only review the history of a chief scientist for AT&T, David Isenberg, to appreciate how much AT&T business school management fears innovation:
even and repeatedly fears to innovate - a symptom of decisions based upon spread sheet analysis.
During the 1980s breakup, AT&T took the business school graduates because AT&T was going to develop new businesses. AT&T MBAs even stifled Unix so aggressively that AT&T's Unix operation was begging POSIX creators to return to System V Unix. Eventually AT&T had to even sell off UNIX.
The product people went with the baby Bells. Therefore the baby Bells develop new and profitable businesses.
AT&T repeatedly tried to buy into new businesses only to sell off those struggling businesses later to mask financial losses. To cover up those losses, AT&T even had to sell off their Basking Ridge NJ headquarters. AT&T needed the money. AT&T is a prime example of how business school philosophies (people who use MBA education) destroy businesses. First and foremost, AT&T is a prime example of how innovation is constantly stifled by those business school principles. Above is only an abridged list of glaring examples. AT&T has been that bad for multiple decades due to their 'MBA school' philosophies - that stifle innovation.