Hacker News Reader

Comment by anderiv | Hacker News Reader

In setting up (and verifying) several VoIP E-911 systems and other PSTN phone systems, I've made dozens of "test" calls without issue. When the dispatcher answers, just tell them that you are testing 911 service and that you would like to verify that you got connected to the correct dispatching center.

I've never had one of the dispatchers become upset at this - they want your 911 calls to succeed just as much as you do.

View on HN

Yep. VoIP contractors (I was one) are usually required to call 911 from every location in a PBX system to make sure that the E911 data shows up correctly.

There is usually a script for this, and the first words out of your mouth should be "This is a non-emergency call. I am testing <insert short description>. Is this a good time?"

It tells the call-taker that they don't need to worry about you, and that you are being careful not to tie up resources. If they need to hang up and focus on priority calls, they will. Most likely, they'll be happy to help.

You only need to do that if you're stupid frankly, all you need to do is insert dialing rules that if 911 call send the BTN/main number, if for all other calls send station DID - often this is done on the carrier side - so you should only need to place calls from one phone for the whole site. It's wholly impractical to do this as well for any site with more than 20 phones, I had customers with sites that has over 1000 stations on them.

Sorry, by "location" I meant site, not station. Yes.

That makes more sense, yes :-D - 911 routing issues were always vexing, usually it was a to the wrong PSAP.

One of the big players had a lovely-insane policy. They'd use your last MSAG-validated[1] address, regardless what new address you had provisioned.

So, user's in Texas, gets an MSAG-validated address, everything's good. They move to Kentucky, change addresses. The new address would have some MSAG issue, and could take days or longer to resolve.

This company thought the best thing to do was to route the call to Texas.

(MSAG is the street addressing system that the PSAPs use. It can be considerably different from the postal address.)

That sounds vaguely like the normal intrado bullshit.

Hey, using Microsoft Streets and Trips is a great way to charge $85/hr to correct addresses.

In 2010 when I was doing it, we'd run it thru google maps if validation failed, similar conceptually.

I can understand doing this as part of your job to test a system but isn't it reckless for people to do it just to test their own phones? You're taking up one of a limited number of phones lines for 10-20 seconds, potentially leaving someone in a life threatening situation unable to get through in the only 10-20 seconds they have the opportunity to make the call in.

If everyone did this at once, it might overwhelm the call center, but in general a E911 call center needs to have far more capacity than there are calls (or otherwise an urgent call might go unanswered).

My understanding is the key rules are:

1) wait for an operator and do not hang up (because if you hang up they assume something bad happened and will send police/fire/etc to you), and,

2) tell them that it is a test call so that they can assess whether they can spend time on you or should hang up to answer a more urgent call.

I'm sure kids call 911 enough as a prank that your test call won't have a meaningful impact on call volume.

In my area there were actually advertisements on the radio asking people to use the lock function on their cellphone to prevent pocket-calls to 911 a few years back. They claimed it had become a problem.

It leads me to wonder if your comment about capacity (which I assumed was the case as well) does not apply evenly everywhere.

Well, if you don't tell them it's a test call and you pocket dial, I would hope that the 911 call center would assume that you're being held hostage and send police/fire/etc. to you.

So, IMO, that the system isn't meant to handle this type of false alarm isn't reflective of call center capacity but rather fire/police/ambulance capacity (and also annoying everyone involved).

The system is built to tolerate 25 people calling in the same accident on the interstate or 100s of people calling in a plane crash. Last specs I heard around '10 was about a half million calls per day country wide. There are about 25 million small businesses in the country (accurate to only about one sig fig). If you assume a business tests a new phone system once per decade and there are about 10K days in a decade (close enough not to matter) then there will be about 2500 test calls per day out of 500000 total calls or basically its a rounding error.

Now on the other hand, don't be a jerk. They're busy at certain times of day and they kinda expect people to make test calls during relatively boring business hours but not during rush hour or at bar close time. Don't be making test calls while a hurricane is hitting the city or while a t-storm is passing thru or a presidential visit or wildfires are kinda nearby, etc.

I've done this helping out as part of PBX installs (admittedly a long time ago) and just be calm and clear and polite when verifying the address routing and remember to say "thanks" and its not a big deal.

> ...there are about 10K days in a decade (close enough not to matter)...

That is quite a bit off, there are ~3650 days in a decade. If you really wanted to use round numbers, would 5K not have been a better choice?

No, because its harder to estimate the arithmetic in my head with 5K than 10K and when I got the result it was so utterly extreme that I felt no need to go back and correct. If it was significant rather than a rounding error then I'd have gone back and rerun the numbers.

Also I forgot to mention I used small biz pbx as the most populous estimate because I assumed j.random.ATT cellphone user wouldn't test, one iphone should do about the same as the next

Reply on news.ycombinator.com