Heavy Data Use Puts a Strain on AT&T Service(nytimes.com)
nytimes.com
Heavy Data Use Puts a Strain on AT&T Service
http://www.nytimes.com/2009/09/03/technology/companies/03att.html?hp
24 comments
Unless it means download from the app store.
Boo hoo. AT&T is offering a product just like any other business. They don't get special sympathy just because they have millions of customers. When I pay my ghastly $100/month bill I expect to get the service I paid for (access to internet whenever I want) and as long as I'm getting that coverage, I'll be happy. Sorry NYTimes, I don't buy it.
AT&T has more of a scalability problem than most businesses. Usually you don't have the laws of physics coming in to interfere quite as strictly.
If they could fix Austin by bringing in some temporary cells, then they're facing something besides the laws of physics, at least for that one town. How does the denisty & bandwidth use compare with Tokyo, Japan?
At some point you have to accept that there is finite bandwidth. Shannon says that if you have an arbitrarily noisy channel, you can get data through it slowly.
The problem is that people don't want slowly, they want it now.
The problem is that people don't want slowly, they want it now.
And how does the red tape required to add a tower/antenna compare?
The iPhone has been out for over two years now. There isn't that much red tape.
Have you been to a community meeting about adding new cell antennas in (for example) San Francisco or Berkeley? It ain't pretty. See for example:
http://www.antennafreeunion.org/
http://www.no-celltower.com/
http://www.antennafreeunion.org/
http://www.no-celltower.com/
Title could just as easily be:
AT&T Doesn't Provide Purchased Service
or
AT&T Networks Oversold and Underfunded
AT&T Doesn't Provide Purchased Service
or
AT&T Networks Oversold and Underfunded
“It’s been a challenging year for us,” said John Donovan, the chief technology office of AT&T. “Overnight we’re seeing a radical shift in how people are using their phones,” he said. “There’s just no parallel for the demand.”
Couldn't they just have spoken to telecoms in Tokyo? Then they could've benefitted from their hindsight.
Couldn't they just have spoken to telecoms in Tokyo? Then they could've benefitted from their hindsight.
Their world view goes something like this:
(subscribers + news subscribers - churn) * average rev. per user (ARPU, like "arr-poo") = $$$
How do you increase profits? A few options:
- Get more subscribers. Costs money in marketing $$$, only helps a little bit. Requires more infrastructure.
- Reduce churn. Customer satisfaction has been in the toilet for some time and will be for the foreseeable future. Multi-year contracts w/ termination fees help keep it under control.
- Sell more stuff to your existing customers (you have lots of those) as long as it doesn't cost you much (SMS, ring tones, ...). Minimal infrastructure requirements.
In this view, crappy phones with expensive additional services like SMS, roaming, ring tones, for-pay WAP, for-pay instant messenger, music stores, push-to-talk, etc., all make a ton of sense. Your incremental profitability is great because you don't need too much gear to support these things and you're already billing the customer. On top of that you can negotiate content deals and pretend that customers care about the exclusivity (mMode, VCAST, whatever).
This stuff is what carriers meant by "data services" a few years ago. What's happening now doesn't fit this model; rolling out better infrastructure costs a lot of money and one person can use a lot more capacity in data than voice. 1000 minutes of GSM audio is only about 200MB and costs about $60, meanwhile the same price gets you 5GB of data. Worse, porting a number is often a big hassle but nobody cares what my IP address is. Imagine a world where your company just moves bits and the government is starting to question your key protection against churn (ETFs). If you're an exec at a phone company this is unattractive. Their attitude is "we don't want to be a pipe" (actual quote from a C-level).
So it's not that they didn't know, they just dragged their feet a little bit. To be fair, AT&T in particular cobbled their network together out of a lot of acquisitions so they probably have had a hard time of it.
(subscribers + news subscribers - churn) * average rev. per user (ARPU, like "arr-poo") = $$$
How do you increase profits? A few options:
- Get more subscribers. Costs money in marketing $$$, only helps a little bit. Requires more infrastructure.
- Reduce churn. Customer satisfaction has been in the toilet for some time and will be for the foreseeable future. Multi-year contracts w/ termination fees help keep it under control.
- Sell more stuff to your existing customers (you have lots of those) as long as it doesn't cost you much (SMS, ring tones, ...). Minimal infrastructure requirements.
In this view, crappy phones with expensive additional services like SMS, roaming, ring tones, for-pay WAP, for-pay instant messenger, music stores, push-to-talk, etc., all make a ton of sense. Your incremental profitability is great because you don't need too much gear to support these things and you're already billing the customer. On top of that you can negotiate content deals and pretend that customers care about the exclusivity (mMode, VCAST, whatever).
This stuff is what carriers meant by "data services" a few years ago. What's happening now doesn't fit this model; rolling out better infrastructure costs a lot of money and one person can use a lot more capacity in data than voice. 1000 minutes of GSM audio is only about 200MB and costs about $60, meanwhile the same price gets you 5GB of data. Worse, porting a number is often a big hassle but nobody cares what my IP address is. Imagine a world where your company just moves bits and the government is starting to question your key protection against churn (ETFs). If you're an exec at a phone company this is unattractive. Their attitude is "we don't want to be a pipe" (actual quote from a C-level).
So it's not that they didn't know, they just dragged their feet a little bit. To be fair, AT&T in particular cobbled their network together out of a lot of acquisitions so they probably have had a hard time of it.
As an aside, my life goal is to have a room like the pictured operations center.
Sorry to stray so far off topic, but:
Miniature computer: http://en.wikipedia.org/wiki/File:IPhone_3G_S_sides.jpg
Minicomputer: http://en.wikipedia.org/wiki/File:Pdp-11-40.jpg
I suppose this is in the same class of usage error as (e.g.) saying 'ordinary bicycle' when you mean a typical one.
Miniature computer: http://en.wikipedia.org/wiki/File:IPhone_3G_S_sides.jpg
Minicomputer: http://en.wikipedia.org/wiki/File:Pdp-11-40.jpg
I suppose this is in the same class of usage error as (e.g.) saying 'ordinary bicycle' when you mean a typical one.
1) It sounds like AT&T badly need to invest in some form of packet shaping for their wireless data that prioritizes cell phone data (calls+text+voicemail) over regular traffic.
2) The iPhone should have the option to use Edge for calls, regardless of 3G signal, if there's congestion. Heck I'd probably choose to use Edge all the time, it's so much more reliable in SF.
2) The iPhone should have the option to use Edge for calls, regardless of 3G signal, if there's congestion. Heck I'd probably choose to use Edge all the time, it's so much more reliable in SF.
Re #1: Calls+textmessages are already completely segregated from data.
I'm calling bullshit on this article. HSDPA (3g data) runs on completely different channels than voice calls. Saturation of HSDPA channels will -not- cause calls to be dropped, poor call quality, or delayed text messages.
It will, as you might expect of a well designed system, instead degrade the quality of other data connections.
That's it.
You'll notice that in the article, no one quoted blames the data usage on dropped calls, except one random 'systems administrator' who owns an iphone.
Thanks for playing, NY Times.
I'm calling bullshit on this article. HSDPA (3g data) runs on completely different channels than voice calls. Saturation of HSDPA channels will -not- cause calls to be dropped, poor call quality, or delayed text messages.
It will, as you might expect of a well designed system, instead degrade the quality of other data connections.
That's it.
You'll notice that in the article, no one quoted blames the data usage on dropped calls, except one random 'systems administrator' who owns an iphone.
Thanks for playing, NY Times.
Shrug. All I know is that 3G = calls dropped, 2G = calls solid (irrespective of displayed signal strength). Repeatable on pretty much any Saturday afternoon in downtown SF.
“Whether an iPhone, a Storm or a Gphone, the world is changing.” Mr. Munster said.
Gphone?? I'm not sure Mr. Munster is exactly a reliable source…
Gphone?? I'm not sure Mr. Munster is exactly a reliable source…
I've been using Verizon' MiFi router and my friends old iPhone with no issue. For me (not much of a phone talker) this works great using text mssge, Instant messaging, email & social networks to communicate w/inner circle. Have Skype on it too, though don't use it too much. Verizon's 3G everywhere router is a bit pricey $60, but cheaper then iPhone's $70 i was paying & getting junky svc.
Is this entirely Apple and AT&Ts fault for trying to achieve a greedy communications monopoly? (Foreigner here, I believe AT&T is the only one offering the iPhone apart from outright?)
There were reports months ago AT&T has started testing home microcells -- that plug into your home broadband, and create a tiny cell zone for voice/etc traffic that's then routed over the internet.
Has anyone seen or tried such a thing?
Has anyone seen or tried such a thing?
Since we're already paying for service, femtocells seem like an insult (unless they're free, which I doubt).
Not only should they be free, in my opinion they should discount my monthly fee if I handle traffic for phones not on my account. That would give people an incentive to set these things up all over the place.
I think I saw a price point of around $100 for the device, with no monthly charge (beyond your existing internet uplink costs).
I agree AT&T should offer them subsidized or even free to help offload traffic from their oversubscribed towers. Still, at a one-time-cost of $100 per location, if it gave myself and others near-perfect coverage at home and work, it would be a bargain.
I agree AT&T should offer them subsidized or even free to help offload traffic from their oversubscribed towers. Still, at a one-time-cost of $100 per location, if it gave myself and others near-perfect coverage at home and work, it would be a bargain.
Shazam runs straight off the iPhone. Its load time has nothing to do with AT&T.