Why It's So Hard To Make a Phone Call In Emergency Situations

antdude writes "BoingBoing reports on why it's 'so hard to make a phone call in emergency situations.' Quoting: '[The thing about] the radios is that they have different sizes of cells. You've got regular cells and then smaller sub-cells. You also have larger overlay macro-cells that are really big. They try to handle you within the small cell you're closest to. But it's a trade off between capacity — they'd like to have lots of small cells for that — and coverage — they don't want to put 100k small cells everywhere. So you might have a cell that covers a mile ara and then smaller cells within that that handle most of the traffic. ... In the end, it does come down to trade-offs. That's true of any network. You're interested in coverage first and then capacity. If you wanted to guarantee that a network never had an outage your capital investment would have to go up orders of magnitude beyond anything that is rational. So each network is trying to invest their budget in ways that make network appear to perform better. The cost of providing temporary extra capacity for the Boston Marathon, that's something that's in the budget and they plan for that event. But when you get something unexpected like a terrorist event, or an earthquake, or damage from a hurricane or tornado, then you have trade offs between capital and how robust your network is. Every time you have an event people say, "Oh, they didn't invest enough." But you look at New York City after Hurricane Sandy and Southern Manhattan was under 6 feet of water — all the buried infrastructure was lost.'"

Comments

Summary?

[Anonymous Coward]

I realize that TFS is a copy & paste job, but WTF? Whomever was quoted shouldn't be allowed to use a phone ever just because they can't speak coherently.

Re:Summary?

[DougOtto]

Yet you type remarkably well. Please pull your pants up.

Re:Summary?

[you-youtube]

I realize that TFS is a copy & paste job, but WTF? Whomever was quoted shouldn't be allowed to use a phone ever just because they can't speak coherently.

Finally, several other countries have implemented all sorts of special procedures for cell phone networks in emergencies (The UK, Israel, Lebanon, and Egypt all come to mind). These sometimes include shutting down cell services once a bombing occurs, but in some of these cases also include using the local version of E-911 as a priority search mechanism for people possibly trapped in rubble after a building bomb or an earthquake, and various other services that mean the system as a whole needs to stay up and function resilently under increased loads. "Common sense" would suggest that the US should have some of these protocols in place too, especially since we have spent literally 10,000.00 % of what some of these other countries have.

pay phones

[schneidafunk]

Pay phones will still work in emergencies. I recall that being a reason for their continued existence in the era of mobile phones.

Re:pay phones

[Anonymous Coward]

The POTS (plain old telephone system, for the young whippersnappers) didn't have unlimited capacity to connect calls either. When many calls were in progress in an area, you could pick up the phone and hear the congestion tone right away. Conversely, if you tried to call an area where many calls were in progress, you'd hear the congestion tone before you'd finished dialing. Only with the internet has it become possible that everyone can talk to someone from a different area at the same time, and only if the ISP hasn't oversubscribed the network bandwidth too badly.

Re:pay phones

[DougOtto]

Ever try to find a pay phone these days?

Re:pay phones

[Anonymous Coward]

Very much beside the point. Only a relatively small percentage of connected landlines can call out of area at the same time. If there is a reason for many people in an area to call or be called at the same time, POTS users experience congestion just like mobile phone users do.

Re:pay phones

[Cramer]

No they do it because it's not necessary. (also "not possible") One cannot build out the PSTN such that every phone can be in use at the same time. The switch itself cannot handle 100% usage, even if it had the trunk capacity for all the lines. Telcos build their infrastructure to meet statistical average and peak usages. Cellular operators can (and do) bring in additional capacity for planned events -- emergencies are far from predictable.

E911 is in the same boat. While there are dedicated trunks to the call center, they do not have infinite capacity. There's finite number of operators to take the calls anyway.

Re:pay phones

[Obfuscant]

The POTS (plain old telephone system, for the young whippersnappers) didn't have unlimited capacity to connect calls either. When many calls were in progress in an area, you could pick up the phone and hear the congestion tone right away.

That's in the days of computer phone switches. In the old days of mechanical relays, there were a fixed, limited number of dialtone generators (and first selectors -- the stepper that handled the first digit you dialed), so if local capacity was reached you just didn't get a dialtone right away.

You still hear this today, but usually after you dial. It's the fast busy signal. The fast busy means circuits are busy, try again. The slow busy means the destination line is busy. If you try a fast busy again right away, chances are good you'll get through, and you'll confuse the person who answers if you accuse them of being on the line when you called a minute ago.

Mother's Day was a big holiday for calling, so it was more likely to hear, or not hear, this happening then.

Re:pay phones

[afidel]

Actually since we changed to packet switched networks it's 100% possible to have 100% activity, in fact today it would be trivial since voice takes so little bandwidth compared to modern networks. To give you an idea 150M people talking to 150M other people would only be 9.6Gbps which can be accommodated by a single peering link today.

Re:pay phones

[Cramer]

In the case of FiOS, Uverse, and cable "digital voice" (where the TA is in/on your house)... if you have an analog phone line, and a GREAT MANY people still do, it goes back to a central office phone switch (ala 5ESS, DMS100, etc.) that has to handle the power and A/D conversion. Those switches cannot handle everyone picking up the phone at the same time. Can they switch that many packets? Sure. Do they have that much DSP (and DMTF decoder) capacity? Questionable and No.

uh, this is common sense

[KernelMuncher]

Anytime you have a large population in a small area all wanting to make calls, the system will be overloaded. Capacity is built for normal use (which is probably 95 or 99% of normal call volume). When there are spikes in demand exceeding this volume, the network will not work as well (or even fail). Also if the network is physically damaged (such as Hurricane Sandy) it won't carry even normal call volumes. How is this not common sesne ?

Re:uh, this is common sense

[rwa2]

Yep, pretty much have cell phone congestion during any large crowd event, such as parades and concerts and demonstrations.

All it would really take is some sort of public education campaign to use SMS in those situations.

911 does take SMS nowadays, does it not? If not, I hear SMS to Twitter / Facebook has been useful for getting people to reach out to their friends for help, who can in turn call an emergency response number... somewhere.

Re:uh, this is common sense

[nine-times]

I think everyone understands this, even if not on a technical level. Anything has an upper limit, beyond which is overloads.

I think the main question in my mind is, what our we comfortable with as a failure of our infrastructure? Maybe we say, "We're ok with the cell phone network going out during an emergency, since those emergencies will be rare and the cost of making the network robust and redundant enough to handle the additional volume isn't worth being able to use your cell phone in an emergency." But then are we really ok with that? If we have a bombing in a major city and people can't really report what's going on because our telecommunications can't handle the strain, is that really alright?

There may be other options, of course. Maybe we want to rethink the design of the cell network to see if we can come up with something than handles the load better and reroutes in case of congestion. Or maybe we just want to figure out a way to prioritize certain traffic so "Important" calls go through while the rest fail. Those things are both easier said than done, but they're other ways to approach the problem.

The problem I see with these kinds of problems is that everyone wants to have their cake and eat it too. They say, "Well why should we waste money building out the network to protect us from a problem that's unlikely to happen?" But unlikely things happen all the time, and when one of them causes a problem, they scream, "WHY DIDN'T WE SEE THIS COMING?" We did see this coming. We decided it wasn't cost-effective to protect ourselves. Pay more attention.

Re:uh, this is common sense

[Lumpy]

"There's been a hell of a lot of money utterly wasted with homeland security since 2001."

Fixed that for you.. you seemed to have made a very common mistake assuming that the money was spent well and not blown on completely useless things.

Re:uh, this is common sense

[afidel]

Not really, for planned events you bring in a few cell on wheels carts per carrier and increase the cell density, this is done all the time for football games and other sporting and political events. Now I'm not sure what the average use rates are for those events, but I bet for something like the superbowl it's well over 50% (for many of the folks at the Superbowl it's more about being seen at the game then it is about the game itself).

Re:uh, this is common sense

[evilviper]

Anytime you have a large population in a small area all wanting to make calls, the system will be overloaded.

That's not necessarily true. The lower the frequencies in-use, the further over the horizon your cell signal can go, and therefore be load-balanced by possibly numerous cell towers.

In a rural area, sure, there's probably only one other tower in range. But in an urban area like Boston, there's tons of cell towers around, which could absorb the sudden spike in demand from that "small area" if properly designed to do so.

Capacity is built for normal use (which is probably 95 or 99% of normal call volume). When there are spikes in demand exceeding this volume, the network will not work as well (or even fail).

If your network "fail[s]" because of traffic spikes, you're doing something horribly, horribly wrong! If that's your "common sesne", I guess common sesne is often wrong...

Besides, acting like cell towers are old telco switched circuits isn't remotely accurate, and doesn't make any sense. A phone call may only need 8Kbps of bandwidth, while with LTE, several customers (at the same time) expect to be able to download at ~50Mbps... That means for each LTE (data) customer you expect each tower to handle, you could alternately handle 6000 voice calls instead (eg. in an emergency, like this one)...

Also if the network is physically damaged (such as Hurricane Sandy) it won't carry even normal call volumes. How is this not common sesne ?

Depends on what you mean by "damaged". For power, on-site generators are fairly inexpensive, and fully automated. Fiber optics are sufficiently water-resistant, so flooding shouldn't knock them out. Stringing lines above ground has been known to be problematic for centuries, so I'd hope the important backhaul is buried, and not affected by storms. And even in the "backhoe" case, these are major telcos, and should be smart enough to have redundant links taking different physical paths.

So if you're talking about those, it's not unreasonable to expect telcos to do a better job, and avoid such "physical damage" in the future. If, however, you mean the cell towers actually being knocked out by winds and flying debris, then I'll concede the point, that significantly reduced cellular capacity is reasonable. However, the former have been the cause of a great many cell outages in the past, often lasting for days or weeks at a time, so I'm generally biased towards assuming most outages are caused by poor engineering, and extreme cost cutting at the expense of public safety, rather than reasonable, practical limitations.

Thanks to the architecture of cellular networks, it's actually practically possible for cells to be MORE RELIABLE than traditional land-lines, though that would obviously be an expensive proposition. As cellular prices fall, though (I'm paying $45/mo for unlimited everything), we should gradually be increasing the standards for cellular phone networks, rather than letting them completely race to the bottom, and cut costs to the bone at the expense of public safety.

(Disclaimer: I may or may not, work or have worked, for one or more major US cellular phone companies.)

Re:uh, this is common sense

[DerekLyons]

Maybe we want to rethink the design of the cell network to see if we can come up with something than handles the load better and reroutes in case of congestion.

You have a vast misunderstanding of how the world works... Congestion isn't caused by bad routing, it's caused by too much traffic in too small an area in too short a time for the available capacity to handle. You can't reroute into or out of a congested area - because there aren't any routes to be had.
 

Or maybe we just want to figure out a way to prioritize certain traffic so "Important" calls go through while the rest fail.

That's been a standard part of telephony for at least half a century.
 

But unlikely things happen all the time, and when one of them causes a problem, they scream, "WHY DIDN'T WE SEE THIS COMING?" We did see this coming. We decided it wasn't cost-effective to protect ourselves. Pay more attention.

Again, a vast misunderstanding of how the world works. Yes, taken as an average and across the whole country - unlikely events happen on a semi-regular basis. But for an individual location? They're still very unlikely.

Re:uh, this is common sense

[geekoid]

Or to get the public to not immediately call to tell everyone you are fine. IN cases like Boston, waiting an hour would have helped a great deal.

Yes, people will be worried, but be a little selfless and give way to people who need help.

Re:sometimes

[oodaloop]

They also said that Marcellus Wallace threw a man out of a window for giving Mia a foot massage.

Post a source, or STFU.

Re:Simple answer:

[Spectre]

Simpler answer: Economics.

Resilience

[Todd Knarr]

Part of it's not just whether the network fails, but how it fails. For instance, in a situation like this the network might be reconfigured to reject incoming calls to the area to keep that capacity free for people calling out. It might start throttling back voice calls to free up that capacity for emergency services and keep the data portion of the network running (and maybe drop the data portion back to 3G or even 2G so it could handle more simultaneous users). You wouldn't be able to call out, but you could still send and receive text messages. And the process for this should be in place. This kind of thing is rare and you can't predict when it'll happen, but it's a given that it will happen so the network operators should have a plan in place for what to do when it does.

And they should also be looking back to Ma Bell's studies on how to staff operators to handle phone calls. They found through a lot of study of real-world traffic that you can't staff for the average volume and successfully handle the calls. Calls tended to cluster, so if you wanted to keep wait times acceptable you had to staff for the peak volumes and accept that that meant you'd have idle capacity a lot of the time. I often get the feeling that the engineering side of the carriers understands this, but the business side doesn't quite grasp the idea of call volume not being a normal distribution.

Re:Resilience

[twisted_pare]

During the Cold War there was a telco exchange in Northern Virginia (I forget the number) that if you dialed through would give your call Federal precedence. It was used by Congress/Senate and high up Federal employees. In the case of a national emergency, those calls would be routed first and others dropped to make way for them. This idea is nothing new. I'm sure something similar exists today with 911 or similar.

Re:Resilience

[guruevi]

Most people think that falling back onto 3G, 2G or 1G would save bandwidth but the opposite is through, the later the generation, the more efficient the data transfer becomes. 4G networks (which are as of yet unavailable in the US) are purely packet-based (voice and data) and can handle much more voice channels over a lot smaller radio bandwidth.

The problem is that the US is quickly falling behind to 3rd world standards on all aspects of society and technology.

Re:Send a text..

[wagnerrp]

That's because you don't squeeze text through. Text messages are placed into what is otherwise wasted padding in the periodic keep-alive packets between your cell phone and the tower. If you are connected, you can send a text message.

Re:Radio

[k6mfw]

CB, and Ham for emergencies. Get it. Love it.

The CB is also a really good way to get real-time traffic updates.

well yes and no. You cannot break a radio transmission like you can with a phoneline or a cell system that gets overloaded. So 2-way radio is good for emergencies.

However, I have been changing my tune about ham radio that its real values are DIY/hacking/experimentation. You still have to pay a license fee (measly $14 to FCC every 10 years) and pay for equipment. You can open it because you own it (cannot do that with many cellphones) and you can modify it as you please (just keep the RF inside the ham bands). And when you have skill and talent to design/modify/implement wireless systems, you can be valuable to those who cannot.

Promoting ham radio only by emergency uses is limiting. Let's be honest, how many disasters occur that ham radio pays a key role? Not many (but don't get me wrong, many public safety officials see amateur radio operators as important resource). So all these people that get caught up in one-day ham cram and take ARES/RACES classes, then wait for the big one.... they get bored and go off and do something else. Emergency planning is important but it is not action-and-adventure where the hero ham parachutes in for the rescue.

CB can be great for traffic updates but for here in Silicon Valley the band is dead. There have been times when the highways backup beyond normal, would be great to call someone couple miles up 101 and ask what the situation is. But this is Silicon Valley and nobody comprehends frequencies less than 800 MHz.

Next week on BoingBoing

[sootman]

Why It's So Hard For a Crowd To Leave a Burning Building Through The Only Exit Doors.

I mean really, WTF?

Re:Next week on BoingBoing

[motokochan]

For those who are having trouble looking it up, Cocoanut Grove fire on Wikipedia.

Re:sometimes

[twisted_pare]

I was there, 200m from the bombs. Phone never had issues sending texts, but could not us Google Voice or regular calling to place a call out. Never had an issue with data/text however, which was useful as I texted folks asking "WTF was that?" Local hardwired wifi never skipped a beat, but sites like Boston.com and Letsrun.com tanked almost instantly.

NCS/GETS

[Hartree]

In fact, this very problem is why there is a US government program that lets certain emergency personnel/offices have priority over normal telephone traffic.

This is also why we don't normally see phone numbers in the 710 area code.

See: http://gets.ncs.gov/program_info.html for an overview.

(Wow, I feel like I'm back on comp.dcom.telecom)

the capacity problem

[nimbius]

gets shit thrown at it from both sides. providers dont feel the need to invest in more towers and users understandably get angry when this problem manifests in dropped calls and network outages.

to curtail the issue, emergency coverage services like COW and COLT (Cellular on Wheels, Cellular on Light Truck) have been bastardized by carriers to augment connectivity for sports events and serve as standby relays during repairs. COW and COLT were designed by the industries to respond to hurricanes and tornados but the allure of having a tower-on-wheels it understandably too budget-friendly for any carrier to pass up. oversubscription and markup are what keep cellular industries alive, just like shared hosting or airlines.
the other issue is as TFA highlights, cellular is just not as robust as say, 25 core ASTRO multi-zone digital radio...arguably because the need just isnt there. if 1 in 5 people cant make contact during an emergency its not a problem, cellphones can be borrowed or the calls can be retried. in law enforcement and emergency services, the PTT button has to work every time no matter what, as a loss of service could result in an emergency turning into a catastrophe.

finally, what i consider 'dark devices' can also create an outage automatically. fire alarms, burglary alarms, and even SIGALERT and some EAS systems (yes, EAS, its cost saving/kickback jack-assery found in flyover states all the time.) for the city/state are critically dependent on cellular networks. in the event of an emergency the activation of hundreds of these devices at once can black out the network pretty fast.

everyone needed to just take a deep breath

[SuperBanana]

When a bomb goes off, you do NOT need to call everyone you know to say "OMG I'm OK!!!!!!" Seriously - the panic is the problem, not the network. Unless you're hurt and need help, put the phone away and keep the airwaves clear for emergency responders - maybe text ONE person and say "hey can you put up on my FB wall that I'm ok?" In fact, go one step further and put your phone into airplane mode and save your battery life, because in a real emergency, charging the phone is going to be a bigger problem. At the very least, disable syncing services. It was amazing how many people thought it was necessary to call everyone they knew in their lives to MIGHT have been running in the marathon or lived somewhere in Boston.

My experience in cell perf after a huge earthquake

[ThePeices]

Where I live, we had this huge shallow mag 6.3 aftershock right next to the city. The cellular networks performed pretty well during the massive emergency call spike dealing with all of the dead and injured, even while dealing with the widescale infrastructure damage that had just occurred. Emergency calls were mostly available in the hours after the quake, and the two main carriers handled the load well all things considered.

We found that SMS messaging was the best method for communicating with friends and family, as voice was under heavy use at the time and best left for emergency use. SMS was good enough really, as it does not require realtime delivery.

The main telco also immediately set every payphone to allow free calling to any phone nationwide, cell or landline for weeks after, and started putting up free WiFi on the top of many payphones. ( The free WiFi is still there today )

Overall, not a bad result from the technology. Good emergency planning can and did save lives.