While everyone is seeing the immediate effect of AT&T planning fixed broadband caps, as a cause for concern, asking if customers may suddenly stop subscribing to broadband, we applaud the decision. It’s not very different from Cellcos using caps to protect their network.
There are two other ways forward for companies like AT&T, and although this doesn’t mean that neither of them will be used as well, this move indicates that it is proceeding down what we see as an enlightened route.
The two other ways are to block traffic from particular destinations and ask for extra revenue from them to deliver it, the type of nonsensical blackmail ISPs have been threatening against companies like Google for so long, which has thrown up the network neutrality debate, or they can randomly traffic shape, that is look at the nature of the data being delivered, such as video and slow it down enough to make it useless. This is what is going on (often in a clandestine, undeclared manner) in Europe and it is making the internet next to useless.
Both of these approaches are to be avoided. On the one hand trying to charge Google and others more, will lead to legal action, consumer uncertainty and disharmony right across the internet eco-system, with the consumer getting poor service, and could lead to old stalwart organizations taking money from innovators. It would be a little like the electricity generating companies asking for some of Microsoft’s millions because PCs can’t run without electricity, and getting it.
The other traffic shaping approach, makes the internet just look like it is running slow periodically, an outcome which would have prevented new services like Netflix taking off, because the performance just wouldn’t be there.
What is interesting is the way the price and cost of delivery comes down and down for the ISP ‘ The DSL chips that gave 500 Kbps, were replaced by chips offering 2 Mbps, then 20 Mbps and will finally be swapped out for chips running at 100 Mbps ‘ but each generation costs the same, for more access network bandwidth. The same is true for DWDM switches in the heart of the core, doubling in throughput almost every 6 months. Routers are the same. So while backhaul remains a concern, every network is getting faster at both the core and the access network and although it means repeated capex, that capex delivers more for the same or less money.
The ISP in turn wants to, in fact is forced by competitive pressures to achieve this for less. A lower $ rate per Mbps, is the key measure of this today. How much do I pay for 30 Mbps. But in a traffic shaped world, that’s irrelevant, and in a world that does not embrace network neutrality, it is also unpredictable. How often are you attached to a 30 Mbps line and only download a file at 275 Kbps? Although right now that is usually down to the remote online resource not having enough bandwidth attached to it, and it being shared by lots of people.
You would like to think that someone like Google or Netflix had done their resource calculations correctly and the problem was not at their end.
But once we introduce a cap, or a meter, although initially there will be no awareness of the cap, it will eventually become meaningful and people will talk about total month cap levels. AT&T is talking about starting off with a 150 GB cap or for U-Verse customers 250 GB, each month. That’s enough for 100 Netflix films, a library of music, and constant html downloads for work, and there is still capacity to spare.
No-one is going to complain about those kinds of caps. Not until HD video and 3D video is commonplace, then they are going to break and what happens then is interesting.
Initially AT&T is talking about charging an extra $10 for every 50 GB of usage on a home’s bill. Again that’s quite generous. How many of us (not so much in the US) have experienced cellular bill shock, because we downloaded five pages of Google Maps on our phone when visiting a neighboring country? This doesn’t result in a mere 10% bill hike, for what might be 50 hours of video, instead it turns a $50 a month, all you can eat data deal, into a $250 monster, because the penalty rates in cellular are so punitive, because bottlenecks in the finite RAN network spill over onto other users. Which is not the case in fixed broadband.
You can see that this kind of cap is sensible and designed purely to give bill shock to the ISP user who is constantly downloading pirated HD movies, and storing them on his or her massive hard drive for viewing over any time in the next five years. The casual or even avid video user will not break this cap.
But data requirements will continue to rise and then the caps have to rise with them. This is where we might see problems. At some stage AT&T will have to make a calculation about what new data level the caps will need to go to, and what they will need to charge extra for that. Because AT&T can then very clearly make a connection between future broadband capex and the extra charging. Or more to the point between future increases in capex and extra charges.
This equation has the opportunity to go wrong down the road. AT&T may take the caps and reduce them while raising the price. But there is enough competition right now in broadband to punish that with customer desertions. Of course if there are less and less ISPs in each territory, that equation will go wrong, and suddenly, like the power company example, you are at the pricing whim of a monopoly. But that’s a different problem with different solutions.
So all in all we welcome the cap, we like the principle of it being high and overuse not attracting a penalty rate, but we’d like to see it stay on track with these principles and that this should not lead to AT&T being reticent to invest in future 100 Mbps DSL technology (see elsewhere in this issue) as it was so very reticent before it made the investment in Project Lightspeed in 2004, which left it in the difficult position of not being able to fight for customers against cable.
AT&T reckons that just 2% of its customers are operating over the cap limits, which must be about right, and it also committed to keeping the process transparent by notifying the account holder when they hit 65% and 90% of their data allowances. Otherwise little Johnny’s secret downloading activities might mean that Dad can’t work at home for the rest of the month. But if non-bill payers are downloading too much, the bill payer will soon be alerted in this way, and only get hit with a small increase to pay, not a befuddling bill five times larger than usual.