Slashdot Mirror
AT&T's City-By-City Plan To Up Wireless Coverage
alphadogg writes "AT&T has created different mobile calling models for every major city in America as it tries to improve a network that has come under fire for poor performance as the data-friendly iPhone has proliferated, an executive said Thursday. Other carriers just use one nationwide calling model to plan for all cities, claimed CTO John Donovan, speaking at the Open Mobile Summit conference in San Francisco. The nation's second-largest mobile operator has had a hard time planning for bandwidth needs in the rapidly changing mobile world, Donovan said. AT&T has seen rapidly growing mobile data usage — and much criticism over its 3G coverage — as the exclusive iPhone carrier in the US. 'If a network is not fully loaded, it's hard to know exactly how much demand is out there,' Donovan said. 'You put all you can in the ground, and they eat it all up, and then you put more in there, and they eat it all up.'" The story notes that mobile data at AT&T has grown 4,932% over the last 3 years.
And if they didnt sign the exclusive deal with Apple, what do you think that growth would have been? Just saying they are complaining all the way to the bank on this one.
Yeah, from the rest of his comment it seems he meant the exact opposite.
"If the network *is* fully loaded, it's hard to know the demand, because you have 100% usage, add more capacity, and quickly hit 100% usage"
"You put all you can in the ground, and they eat it all up, and then you put more in there, and they eat it all up"
This is the typical, in this case subtle (but in other cases not subtle) blaming of the consumer for overusing network resources beyond some mythical "reasonable/predictable" amount that service providers cling to in rationalizing their retarded infrastructure expansion plans.
News flash: your network and every other corporate network is at capacity already and you're overselling subscriptions. Don't add one tower and then complain that those data-hungry fiends are using the new bandwidth so quickly. Either think big and grow some balls about expanding your network, or quit complaining and admit that you've resigned to mediocrity.
Take any area on Earth where you are not at max capacity and then model data usage per phone. Done. In what way is this difficult for a multi-national megacorp?
This guy's quote is BS, if you as the owner of your traffic don't know how much demand there is either by system monitoring and/or usage patterns for specific type clients (with demograhaphics tagged along with it, because ATT sure as hell knows its clients profiles and/or can buy such data from 3rd parties) then they need to get out of the business. Either way ATT has slacked on its network, let Verizon (good for them) to compete and do it well and then blame poor performance and oversell on its lack of knowledge. That is just BS, they know, don't care until it hurts in the pocket... And exclusive contracts with big hardware vendors does't help the public, its own customer base, as well as its image. Shame on ATT.
'If a network is not fully loaded, it's hard to know exactly how much demand is out there,' Donovan said. 'You put all you can in the ground, and they eat it all up, and then you put more in there, and they eat it all up.'"
You've never done any kind of network administration, have you Mr. Donovan? You designed your network for average use, not peak use. As anyone who designs networks for a living will tell you -- it will function perfectly well until it reaches close to or at 100% utilization, at which point it'll choke and die horribly. Had you excercised proper engineering methodology, you would have known to test each product/application being put on the network in test markets and used the use data to predict what the peak would be, and then only deploy it when you had a 20-50% greater capacity than what the data suggests.
But alas, you eschewed best practices to save a few bucks -- all those profitable quarters and executive kickbacks, all the while your towers were backhauled on 512kbit DSL and fractional T1s. Your infrastructure's been rotting for a long time, sir, and the iPhone has nothing to do with your failure as an executive to execute a proper deployment plan that accounts for growth. You should be ashamed: The chinese mobile phone network has over 500 million subscribers, and their plans are cheaper, have better options, and their infrastructure is far more modern. China has similar problems to the United States in terms of rural development and rugged terrain for deployment -- and yet you've abjectly failed to not only do your case studies, but even do exploratory research within your own market.
It's amazing that this level of incompetence is rewarded by our society.
#fuckbeta #iamslashdot #dicemustdie
The exact same observation is made with highway construction--but it has led transportation authorities to the opposite conclusion: if the more you build, the more people use the resource--then clearly the answer is to not build any because you'll never fix the congestion, and you'll just encourage more people to use the resource.
You've never done any kind of network administration, have you Mr. Donovan? You designed your network for average use, not peak use.
I think what he is saying (badly) is that you can't find the peak if your network is constantly at peak.
Not to mention it's hard to figure out what the real peak will be in a few years with 4600% growth in average use.
To put it simply, that level of growth caught everyone flat footed because people just did not use data plans that heavily before. AT&T is still trying to figure out how to adjust. I know they are building out because service in my town (Denver) has improved, but obviously they are struggling to have improvements match growth rates. And they probably will for some time... hopefully they are starting to realize they need to lay down new network capacity at a far greater rate, whatever that takes.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
It's a telco monopolist mentality.
Gosh, we build, then it's at capacity. That's not what we had with landlines!
Remember: this isn't the AT&T of old, with wizened scholars. This is Southwest Bell that sucked up the other Baby Bells, then chose GSM as their infrastructure and got in over their heads. They're still clueless as to what success they've had as a result of Apple's business models. Apple, OTOH, could have 5x the customers if they simply shipped a (w)CDMA/GSM world phone.
---- Teach Peace. It's Cheaper Than War.
...offer something that is your core competency and people want to use it. This is what AT&T DOES. They sell data transit. Move the bits from here to there. That's their job. Their ONLY job. And they suck at it. If that's not a damning indictment against monopoly/duopoly, I don't know what is.
If only my employer had this problem. People beating down their door to obtain the service they most like to offer. Oh what a terrible fate. And they fumble it and practically fail at it. It would be unbelievable if I didn't already know they're more than half of a duopoly.
Worse, they got a giant government handout a decade ago, before giant handouts were quite so common, and they STILL fail. How much of the free broadband money (tax credits) did AT&T née SBC née Bell South née AT&T get from the gub'ment? Billions. And here it is 2009 and they're having trouble moving data from a few million phones. It's crap like this that is the reason why Third World countries tend to stay Third World. We truly have become a banana republic, and we're going to destroy ourselves if we keep up that habit. It does not work.
If he'd been doing case studies from the start and deploying products in test markets on a schedule
Come on, how are you going to "test market" something like the iPhone? It releases everywhere, all at once - and it's been a rollercoaster for them since.
I have seen so many test market plans from large companies go up in smoke. Invariably there is something large no-one accounted for in the testing. You can test and test to some degree, but there are always at least a few huge surprises when you deploy - and the growth of actual data use on hand-held devices is simply something that could not be "tested" for as it also goes hand it hand with application growth, which would not occur at the same rate in test markets.
But they disregarded that in order to release the iPhone nationwide all at once as part of a huge marketing campaign,
But if you want to blame someone for that, it's all on Apple - not AT&T. Honestly though, I don't see that plan really working for them well even if they had been able to do it - as I said before part of the rise in bandwidth is the rise in the number of applications that make use of that bandwidth.
And honestly, who even does small market test releases anymore? Did Verizon do that with Droid? No, because you simply cannot take that kind of time in todays mobile market.
The problem is, it costs more to rush a deployment after a problem like this presents than to do a phased upgrade, when there's opportunities to cut costs by careful selection of distribution channels, contracting during periods of reduced production, etc.
Ah, but this careful approach also would lose a lot of customers, and the iPhone would not have been a success had it been tried as it would have given everyone else way too much time to catch up.
Personally as a customer I'd rather have a network with a few issues and an iPhone than be wondering when it would come to my town... that goes for any device, people always bitch when you release it into a limited market first. Phones are not a thing you can do that with, although it works for other tech sectors.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
...and there'd only be 50 cell phones per city, because no one else could afford one. (Back when AT&T was a monopoly, this was actually the business model: Cell phones were planned as a high-end luxury item found only in limousines and such.) For guaranteed 100% availability of "what users are paying for," every cell site in the network would have to have one radio channel per customer... and that channel would sit idle and unused 99.9% of the time, but it still has to be there, ready and waiting, in case all 50 customers are stuck in the same traffic jam.
To borrow a downthread commenter's analogy, this is the difference between paying for 1.5Mbps DSL and running a full-bore T-1 line to your house: One of these things costs 5x more than the other, and if we insist that ISPs build to that capacity level then most of today's customers will be priced out of the market.
Not that AT&T couldn't have done a better job of growing their capacity, but they really do have a tiger by the tail here: The iPhone is hammering their data network like nothing before or since. Even if they had accurately forecasted the iPhone's impact, they wouldn't have been able to build enough capacity in time.