Comments on news posted 2012-08-17 08:58:14: Earlier this year AT&T began talking about their latest network neutrality violating money maker -- the idea that they could start charging app makers a toll if they wanted to bypass AT&T's caps, thereby getting a leg up on the competition through "p.. ..
there oughta be a law... an idea for lawyers from the FCC/DOJ with an itchy lawsuit finger to remedy real quick. infact just ban usage based byte charging altogether.. and be done with it.. along with state laws blocking 3rd part ISPs from wiring communities w/o 21st century broadband.. carrot & stick.. call it the IF you don't build it, someone else can, and you have NOTHING to say about it, in court or otherwi$e (corruption lobbying).
Depending on how much this "toll" is the app developer could pass it along to the customer and then BAM FoxFI unlimited tethering through your phone with no caps. Or build a browser into the "premium" app and all of a sudden you got unlimited data.
I don't think AT&T thought this through very well.
They will charge the app developer by the drink. So if you go unlimited, then they get slapped w/ a $1000 usage bill.
The point is they are equating this to the old telephone model which has some merit. So dial direct, then you pay. Dial 1-800 the business pays. The only problem is that the phone model doesn't jive w/ the internet because in phone land the 800 model is for customer relationship and the usage of the phone is limited to the interaction. In the case of an app the relationship is constant, so the costs would be enormous.
Add on top that w/ 800 the business could use any provider (to lower cost), ATT is suggesting that they are the only toll and hence you pay their rates, period. This is circa 1970, the good ol' days of T, just before the breakup.
Since there are no government regulations for weights/measures how is anyone to trust T to meter correctly. This begs of caps today which are just begging to be regulated so carriers can't toy with people.
This isn't the old circuit switched days--and not black and white.
Even if it's technically possible, I seriously doubt it gets off the ground unless AT&T is able to scoop some seriously big players, e.g., Google, Facebook, etc. I doubt the companies AT&T needs on board to make the "troll toll" effective are willing to play ball.
I do no see the violation. Can the subscriber use the application to access the network without a deal in place? I assume that the answer is yes ... but the traffic counts against the users cap. In the case where an app vendor (or web asset for that matter) enters into an agreement, the traffic would still flow ... just not included in your cap.
As long as there is a cap, and this is not the discussion, this business mechanism seems fine.
AT&T is trying to revive a long dead business model. 800 calls made sense when toll calls were expensive. Nowadays, with distance insensitive buckets of minutes, who cares about an 800 number? They think framing the charge that way will make it easier to swallow. They're just trying to keep the ghost of Ed "Tubes" Whitacre alive.
Since there are no government regulations for weights/measures how is anyone to trust T to meter correctly.
Just like you trust the carrier to bill voice seconds/minutes correctly. Some carriers count from first ring, others count from call pick-up. Some close the line quickly after the call ends, some let the call 'hang' for a few seconds before closing the line. There may be several seconds difference between how long your calls have been and how long you got billed for.
I guess it depends on how traffic bypasses the meter. It would be fairly foolproof if traffic bypasses the meter when a "premium" app uses a known set of network end points. However, if it's not tied to known end points and it's just the app identifying itself to the meter, how long will it be before hackers root the phone and send random "premium" app signatures, regardless of the app in use, to bypass the meter? Granted, it's a bit dangerous since they know who used the data but if you own the premium app, whose to say you just didn't use it a lot? Of course if they can track the end points it could be a dead giveaway when lots of traffic was exempted from sites the app wouldn't use.
The premium app could protect itself by passing on usage costs to consumers but then what's the point? That's like calling a free customer support 800 number and then getting a bill in the mail for the long distance charges. You might as well just pay AT&T unless the company heavily discounts it.
at&t and Verizon have data plans with caps. Now data is data obviously. But so is messaging in fact so is making phone calls. It's ALL data. Yet they don't charge you for that data. At least not as data. So somehow at&t already has the ability to tell that phone call data and messaging data shouldn't be counted as data.
On capacity-constrained networks, usage based billing makes sense. Unlike a fiber or DSL network, where capacity upgrades are relatively inexpensive relative to the cost of the initial plant (and the initial plant in the case of DSL has a technical ceiling), HFC and cellular systems have incremental upgrades (more bandwidth, smaller cells/nodes) that can be tacked somewhat easily to expenditures. It's far from perfect, but especially on cellular networks (with relatively limited capacity) the cost of a few super-heavy users may actually end up being a whole new cell site, if you're in an urban area where capacity is actually constrained.
Now you can choose to base your network infrastructure and user pricing on average usage rather than tinkering with pricing to monetize the users who, indirectly, cost you the most money in plant upgrades...Sprint does this with smartphone users...but if the pay-per-byte charge is reasonable (and it isn't on AT&T and VZW right now, but at one-third of current rates it would be) you've got a situation where usage based billing works out.
Again, current UBB rates are too high. $10 per GB on mobile wireless is too high. 20 cents per GB on DSL is too high...DSL probably shouldn't even have caps. 20 cents per GB on cable, billed in $10 increments, doesn't have as much granularity as it should. However if rates were $5 per GB, billed in 100MB increments, for mobile, 50Â¢-$1 per GB billed in gigabyte increments for fixed wireless and 10Â¢-20Â¢ per GB billed in dollar increments for cable, that's reasonable use o UBB. For fiber or DSL providers, you don't need caps as long as you make sure that someone isn't reselling the connection.
Then again, I'm pro-UBB compared to everyone here. Maybe because, if I started a wireless ISP (and there's a good chance that I will), I'd have a $25 plan with enough usage and speed for 40% of people (2 Mbps down, 1 Mbps up, 50GB per month probably, with reasonable overage charges), a $50 plan that was good enough for 80% of folks, and on up until nearly-unlimited (probably 1TB) for around $100, delivering as high a speed as I can do reliably over fixed wireless and charging overages in small enough blocks that, bove the $50 level, you'd be paying for usage like you pay your utility for electricity, rather than like you pay your cell phone copany for overage minutes (though the concept of overage minutes is quickly going away).
Phone calls and messaging don't go over the data network yet. They're all circuit-switched. When VoLTE happens this will change, but by then AT&T won't have to worry about billing for minutes or texts; it will all be unlimited.
I think I need to finish the blog post I was going to write on this awhile back.
AT&T can basically sell app developers "free to user" traffic to a certain IP address or range. Then, if a user access an app that uses that connection, the data used gets billed to the app developer instead of the user. Kind of like 1-800 numbers.
As for provider choice with 1-800 numbers, I guarantee that AT&T will allow "toll free traffic" providers to resell services, just like 1-800 providers provide various packages for their services. In the days of 1-800 numbers, the 1-800 provider still had to pay the originating telephone company's long distances fees; they were just billed at a different rate than standard customer-billed long distance. If I called a 1-800 number from a Hill Country Telephone Cooperative line, HCTC would bill ABC 1800 or whoever for X cents per minute for the call. Likewise, AT&T would bill the toll-free data provider for any data transferred between their mobile network (and I guarantee that this would at least start as mobile-network only; WiFi hotspots would be exempt) and the content provider's IP space.
This is a win for everyone (believe it or not) if AT&T offers a standardized rate structure to everyone asking to do toll-free data, based on the volume of data that they transmit. My bet is that AT&T won't want to deal with anyone transferring less than 1,000 GB per month to their subscribers, but at that point they'd be willing to give some nice discounts compared to the $15 per GB that theyll charge for overages. Heck, the rates will probably even be nice compared to the rates on their 20GB plan. I'd expect $3-$5 per GB depending on volume per month...low enough that CDNs and other such providers could resell services to smaller app developers and still leave those developers with rates lower than their customers would pay to AT&T directly. I this case, it's all about billing simplicity...it's easier to collect $5000 per month from one account than $8000 per month from 800 accounts.
Also, most content providers won't need to do this. Facebook has worked with some cellular providers to do what they call "Facebook 0" (no cost to the subscriber for a very stripped down version of FB's mobile site), however last I checked Facebook doesn't do a ton of video streaming, and it's in their interest to transfer as little data as possible anyway. Point is, unless you're pushing tons of photos to their site via your phone (or videos for that matter), you're using very little of your monthly bandwidth cap on Facebook...though they might participate in this program anyway.
By contrast, Netflix (and other streaming video sites) use a lot of bandwidth, and they're parties that AT&T would love to bill from a content provider perspective. They'd probably offer some nice rates, too...think a few dollars per gig...money that AT&T can use to get their LTE network built out just a little faster, or with a bit more capacity. Users would probably end up with the choice of paying for Netflix streaming with their own mobile data plans, or picking pay-per-view so that Netflix pays for the data, then bils the Netflix customer. Of course, Netflix will fight this tooth and nail because that decreases the number of users who are willing to pay for Netflix, but at that point Netflix should just start recommending that everyone go buy Sprint phones...if Sprint's okay with that
One huge issue here is Net Neutrality, of course. AT&T should not be able to provide any content whatsoever that is exempted from its own data plans, because that's a common carrier conflict of interest; AT&T would be able to subsidize its stuff with fees paid by its competitors. But if AT&T can keep its hand out of that cookie jar (and this is something that can...and maybe should...be regulared), the "toll free data" model works for mobile. Though if AT&T tries it for DSL, customers should vote with their wallets.
I this case, it's all about billing simplicity...it's easier to collect $5000 per month from one account than $8000 per month from 800 accounts.
No app provider is going to trust a $5,000 charge without having details (who, when and quantity). Nothing easier here.
The carrier still has to bill the consumer for their plan and other fees. That money still needs to be collected whether it's $100 or $110 ($8000/800 = $10/user). Nothing easier here.
If I was a carrier, why wouldn't I just collect the $8,000? Where is my incentive to give away $3,000 in profit?
<cynic> Does Mr. Carrier look better if they hide the fees behind someone else? For instance, if NetFlix charges me $30 a month for mobile streaming so it can cover a $10/month charge from Mr. Carrier, he still gets his jack but it's hidden. Next year Mr. Carrier gets to raise the data rates and NetFlix gets to explain why I have to pay $35 every month instead of $30. NetFlix will claim their costs increased but Mr. Carrier is smart. The fine print says that NetFlix cannot divulge the details of their deal. Information can leak but with no hard evidence, the money is still going to NetFlix and they are forced to justify it. Plus, Mr. Carrier doesn't have to deal with angry customers who balk at seeing overage fees on their bill. </cynic>
Re: Technically possible...and decent if done right
IP tracking etc. is easy; AT&T can already do data by session, and doing data collection by source/destination is just one additional small step. If content providers want fine-grained information, AT&T can give it to them (even though the content providers could just as easily do such audits from their side...actually it would be easier to do it on the content provider side).
On the "only their content" issue, my guess is that AT&T would have contracts with the upstream data providers rather than the app makers whose content isn't being pulled. Also, remember that this sort of thing only makes sense on data-intensive applications; people aren't going to go over their caps pulling down maps. On the Akamai side, I guarantee that there will be certain Akamai IPs where Akamai will bundle CDN services and toll-free data into one product, if this catches on. Same for other CDNs.
As for why the ISP would leave money on the table, they wouldn't be. Billing efficiencies are one reason. Another is that cheaper data will stimulate demand, possibly to above (in revenue terms) what they would be if a user had to pay for the data directly (at a higher rate). Call it the Southwest Effect of data.
No really, why do I have a cap on using my phone where I pay for data? So now they want to get paid for data delivered and data received?
Nope that's not anything like what they are talking about.
It's like this. Right now using Netflix over mobile uses data. Now if at&t could charge Netflix a fee( I assume to be in the millions ) then people could use Netflix over at&t network and it won't count against the cap. Of course you can bet Netflix would pass this fee onto it's customers.
*shrugs* It all comes down to how a private corporation wants to run their network and allocate scarce resources (spectrum) to subscribers. Toll-free data, where the content provider pays, is one way of doing this in a net-neutral way (if it is done according to certain specific parameters). Remember that telco-grade network equipment for licensed spectrum is expensive...
At the backbone level, sure. I can get a 10-gigabit backbone line for $7500 now, whereas it was $10000 under the same terms last year. However that bandwidth is delivered to a data center in Dallas. There's a lot of middle- and last-mile gear between that point and the subscriber.