[besson3c]

It looks like there is a ton of FUD and misinformation about all of this. I found this article to be the most sensible in sorting all of this out:

AT&T vs Unlimited Plan Users: Round 2

MyWi uses the same APN as Mail, Safari, and everything else so I don't buy that AT&T really knows that you are using this app. What they are looking for is simply patterns in data consumption. According to this article only users with unlimited data plans are getting the text message from AT&T.

[Cold Warrior]

By looking for patterns that indicate tethering, that implies mywi and others. att doesn't care about the particular app, just that they're not getting paid. The app passing data to their apn would be invisible to them anyway, so they'd have to look for indications. That article you linked to lays it out well I think.

[brassplayersrock²]

Originally Posted by imitchellg5 

Virgin will have coverage if Sprint does.

Out of context this would sound very wrong.

[besson3c]

Originally Posted by Cold Warrior 

By looking for patterns that indicate tethering, that implies mywi and others. att doesn't care about the particular app, just that they're not getting paid. The app passing data to their apn would be invisible to them anyway, so they'd have to look for indications. That article you linked to lays it out well I think.

But they have no way of looking for patterns that indicate tethering if you are using the same APN, just high data usage. Yes, they can see if you are hitting OS X Software Update like the article suggests, but I highly doubt this would be a blip on their radar if you kept your data usage in check. There is no way they could be sitting back and monitoring all traffic patterns regardless of usage, that is just highly inefficient.

As the article suggests, they have gone after people that have been tethering and not tethering, there have been several false positives.

[Cold Warrior]

OS X or Windows SU (or DNS lookups to similar services) and Flash/content are indicators regardless of APN. They don't have to do this in real time -- they could just scrape the logs through some filtering criteria:

user_type:iphone | plan:unlimited | dns_lookups:indications_host_list.txt > tethering_sms_recipients.txt

[besson3c]

Originally Posted by Cold Warrior 

OS X or Windows SU (or DNS lookups to similar services) and Flash/content are indicators regardless of APN. They don't have to do this in real time -- they could just scrape the logs through some filtering criteria:

user_type:iphone | plan:unlimited | dns_lookups:indications_host_list.txt > tethering_sms_recipients.txt

Inspecting each and every DNS packet and logging these requests is ridiculously expensive, which is why DNS is generally not logged. They would be best off monitoring actual traffic to IPs/subnets. I'll give you the OS X/Windows SU being a giveaway and the Flash with iPhone users (although you can run Flash if you jailbreak your phone), but even then they'd have to keep track of which sites don't offer an HTML5 alternative to Flash video and stuff.

The point is, there is no smoking gun evidence, and they've been getting a lot of this wrong.

[turtle777]

CW, exactly.

Really, to believe you could fool the ISPs is foolish.

If they wanted to know, they could. They can detect with very high probability that your use is NOT just on a smartphone.

Heck, they don't need to filter everybody's data. They just need to start with the top 1% bandwidth hogs. I guarantee you, the top 1% will be 99.9% tethering.

Send them ONE huge overage bill, then offer to settle for 20%, and move on to the next 1%.

WAY cheaper than investing in new infrastructure.

It's beyond me why it is taking them so long to double their bandwidth by just killing the top 5% of the data hogs.

-t

[besson3c]

Originally Posted by turtle777 

CW, exactly.

Really, to believe you could fool the ISPs is foolish.

If they wanted to know, they could. They can detect with very high probability that your use is NOT just on a smartphone.

Heck, they don't need to filter everybody's data. They just need to start with the top 1% bandwidth hogs. I guarantee you, the top 1% will be 99.9% tethering.

Send them ONE huge overage bill, then offer to settle for 20%, and move on to the next 1%.

WAY cheaper than investing in new infrastructure.

It's beyond me why it is taking them so long to double their bandwidth by just killing the top 5% of the data hogs.

-t

Well, evidently they can't, if you read the article, not with any accuracy.

[turtle777]

Originally Posted by besson3c 

Well, evidently they can't, if you read the article, not with any accuracy.

What ?

The article doesn't say at all that they can't.
Because clearly, the article lists several ways how ATT could tell if tethering is involved.

Seems like ATT is not even bothering to prove you're tethering. At this point, they just threaten the bandwidth hogs, and hope that some will stop or leave.

I say Fugg it and just kick out the bandwidth hogs.
They're not losing anything of value.

-t

[besson3c]

That's the point, it's the bandwidth that's the problem, not the tethering if you keep your usage in check. The article shows some possible ways for them to determine whether you are tethering, but there are always ways to avoid the detection, these methods are crude and imperfect, and chances are you will go unnoticed regardless if your usage is minimal.

[Cold Warrior]

That's a lot of faith in att. You mentioned it's impractical to log all dns -- at an ISP level, I'll agree. But think how small a pool it is to take your unlimited iphone subscribers and log their activity just a bit more. Even in rolling pools, and even then only samples not 24/7 logging. I think you're underestimating the ease by which this could be done. We're talking about the same company, right?

[besson3c]

They definitely target certain demographics, including the unlimited users. My main point I was trying to make was that moderate usage may not be affected, and that there is a lot of misinformation circulating