Tracking traveler phones lets Cincinnati airport cut security line wait by third | Hacker News Reader

Tracking traveler phones lets Cincinnati airport cut security line wait by third

theverge.com

60 pointscleverjake40 comments

Can't help wondering why they have the "choose the line" step. An obvious improvement is to have one line and direct passengers from the end of the line to the next available station.

Then again, a better solution is to staff all screening stations all the time or build more. Why do we wait passengers wait for a minimum wage screening person? By do this we effectively say the passengers time is worth nothing. Their employer might no appreciate them having "idle time" with no-one to screen but I'm sure the passengers would sure like it. We don't mind firefighters and paramedics having down time, this should be extended to airport screeners.

I totally don't understand this, I would presume each line is processing people at 100% (of their capability) - how does controlling what line a person gets into going to shorten aggregate processing time unless some processors are sitting entirely idle, which we all know would never happen.

I'm pretty sure there is some fundamental queueing theory to back this up, as it's the main takeaway I had from a queuing theory course I took in college. Can't remember the details though, so I'm not sure it would still apply if a single queue entailed long walks from the head of the queue to the nearest available 'processor'. EDIT: more background here https://possiblywrong.wordpress.com/2010/12/26/why-multiple-... It appears you might be right in practice.

I was going to link that too but I realized it had total throughput of cashiers as faster than customers. That means there are idle lines without people switching, which is not realistic at all.

I'm really at a loss about what they made faster.

From your link:

"for 3 cashiers, it’s about 3 times faster than having a single line for each cashier.”"

I would very much like someone to take me to a place in the physical world and show me an improvement of this magnitude. I simply cannot see how (in a setting like a bank or grocery store) this theory is nothing short of absolute insanity.

EDIT: I just watched that video - is that guy trolling viewers? If you lose 1/3 of your cashiers, it has no effect on total throughput if you are using a single line queue?

Just wow.

Probably not helpful, but after reading this I can't stop wanting to make a device that appears to be ~20-30 devices by randomizing its MAC address quickly and causing extra staff to be sent to my line any time I was at a place that used a system like this.

You wouldn't need a device, would you? Just a rooted Android phone and an app. Maybe a custom kernel.

You're not getting staff directed to your line, you're just screwing with other passengers who are fed erroneous data:

"The system allows the airport to display to passengers how long the wait in each of its security lines is."

All that system does is show you (ie passenger making decision where to go) which line has fewest people in it.

"The system allows the airport to display to passengers how long the wait in each of its security lines is. It also allows the airport to adjust its resources to respond to demand."

I feel that this supports the GP and contradicts your statement. If you're sure the article is wrong, can you support that?

I guess this is a matter of interpretation.

My read is that they simply show accurate line wait expectancy, thus ensuring that all lines are utilized (as opposed to one line being empty and the other having a 10 minute wait).

From a business perspective it's much simpler for an airport / TSA to put same staff numbers across and expect passengers to figure out optimal route, as opposed to making staff move around.

There's nothing erroneous about reporting random mac addresses. I think that should be the default behavior for devices. Mac addresses shouldn't be used for tracking. If the airport reports valid data in an incorrect way that's their fault.

If you read my comment carefully you will see that I did not express the sentiment that you are arguing with.

My point was that the parent comment action was primarily directed toward fellow passengers (ie it's a dick move towards them, airport staff aren't going anywhere :), not whether 'mac tracking' is right or wrong.

Random MAC is fine. Pretending to be 20 devices is obnoxious.

No, it's not. One is free to preserve their privacy by using 20000000 if not 20 device MAC addresses. The aim is not to be tracked - which would imply changing the MAC address regularly.

Regularly is fine. Maintaining 20 simultaneous MAC addresses screws up things like DHCP and QoS.

The security theater is obnoxious. This is just gaming the security theater.

Measuring line size has nothing to do with security or security theater.

The fact that the lines are as long as they are has everything to do with security theater.

This is exactly what Pry-Fi [0], an app for rooted Android devices, does.

[0]: https://play.google.com/store/apps/details?id=eu.chainfire.p...

When iOS 8 came out Apple announced a MAC addresses randomization feature to prevent device scanners like this one. But after people tested the feature it turned out that the scheme only works under a very narrow set of conditions that makes it nearly useless. I haven't heard if this has changed in 8.1 or 8.2 but hopefully Apple will move to stop these shenanigans, as well as Google's Android.

Until that happens, you can prevent this kind of tracking by turning your WiFi radio off. For battery life purposes, it's wise to do this anyway when you're not in a location where you intend to be using the WiFi connection.

Are there any local-only contextual apps which can turn wifi on or off based on GPS location?

Llama is really nice.

I like Wi-Fi Matic (https://sites.google.com/site/wifimaticapp/). It uses the phone network cell identifiers instead of GPS (so it works even where you can't get a GPS fix), and auto-learns the cell identifiers which can be seen in the range of your wifi networks. And its source code is available (I originally found it on f-droid).

thanks, installed.

It doesn't look like MAC randomization is relevant to this scenario unless you're cycling the device on and off the network while you're waiting in line.

What it would do is prevent them from tracking you at different locations or trips, although obviously the TSA & airlines still know you were there.

It would be cheaper, easier, and quicker to just do away with security theater completely. Pesto, no wait time at all!

A bit surprising that a sufficient proportion of people leave wifi and/or bluetooth services on while at the airport, for this system to get a large enough sample size.

Why would it be surprising? I only turn those off when I get on the plane.

I can't imagine that tracking cell phone emissions is the simplest way to estimate the number of people in a line.

I also can't imagine this is the only thing they are doing/planning to do with the data.

Maybe more significantly, they aren't the only ones who are collecting this data. A lot of retailers are doing it, and apparently some governments and others are doing it too, all to keep track of people's whereabouts and habits.

There are also people who get in a position to map between a device MAC address and an online identity. The easiest case is people who operate hotspots that make people pay; there are also hotspots where you can authenticate (sometimes optionally) with an account of the hotspot operator's service, which may now include Facebook, not just mobile carriers.

In prior mobile OS versions -- at least -- some mobile apps were also directly collecting the wifi MAC address and sending it back to the app developer.

It is the cheapest to implement, commodity $4 wifi/bt chip is enough to gather all the data, and you dont even have to process it all that much, 200-400MHz router SoC can do it without external controllers.

> looks for devices over Wi-Fi and Bluetooth

There must be lot of devices which are not connected to these two. So, this makes their system not completely accurate.

They could have used other options like - 1. Sensors on the floor 2. Sensors on the tables where we put our stuff 3. How busy is x-ray scanner belt? 4. How fast people are getting through scanner door.

In all of your suggestions they would have had to pay for the sensors themselves.

By using the passenger's phones, they're essentially crowdfunding the cost of the sensors.

In both cases, they still have to pay for the "collector and analyzer"

Some people may even switch off their phones when they will come to know that they are getting tracked.

Another reason for their system's inaccuracy is - not many people use airport wifi. Most of us have data plans.

> Some people may even switch off their phones when they will come to know that they are getting tracked.

They don't need to count each and every person to determine relative wait times, as long as the portion of wifi-enabled travelers is roughly the same in each area.

> Another reason for their system's inaccuracy is - not many people use airport wifi. Most of us have data plans.

You don't have to be connected to the airport's wifi for this system to work. The phone's wifi radio just has to be turned on.

For wifi at least, it doesn't have to be connected, the radio just needs to be on.

More advanced systems are build around SDR platforms and passively scan GSM band for TMSIs.

I don't mean to get all tin foil hatish, but why on earth do they need things like MAC addresses to count how many people are around? Is there not software that could take video of the area and count these people? This just sounds like an excuse to be overbearing...

I'm wondering if the rapid decrease in the number of flights out of CVG might have also caused the security waits -- since Delta left the Cincinnati airport once the Northwest merger was approved, they keep on cutting flights back...

Reply on news.ycombinator.com