Will 3G or WiMax fix the slow iPhone?

We all recall the carriers’ pride regarding 3G data speed capabilities. It was easy to make big statements while hiding behind poor man-machine interface that has been greatly discouraging the use of portable devices for real Internet access (streaming video and music etc.). The iPhone’s interface is so good that we can really enjoy the cool stuff offered. Could we? OK, not so fast …

David Pogue from NY Times is telling us: “The biggest problem is the one that most predicted: the EDGE network Apple chose for the first iPhone. Unless you’re in a Wi-Fi hot spot, Internet browsing is going to be painful, Pogue said. “The New York Times’s home page takes 55 seconds to appear; Amazon.com, 100 seconds; Yahoo, two minutes. You almost ache for a dial-up modem”.

John Markoff from the New York Times: “Early reviews of the iPhone, while positive, have faulted the slower network because it will limit the palm-sized wireless computer’s utility in making the Internet easily accessible on the go.”

W. David Gardner at InformationWeek: “The early reviews of Apple’s iPhone are in and, while there is near-unanimous consensus that the handset is wonderful, most reviewers point to a major weakness: the slow AT&T network on which the iPhones will operate.

The Edge (Enhanced Data Rates for GSM Evolution) network can be painfully slow; it generally operates at between 75 and 135 Kbps. In a broadband world of users accustomed to the instant gratification of 600 Kbps to 3 Mbps, that slow speed will frustrate many iPhone subscribers.

The iPhone can switch seamlessly from Edge to a much faster Wi-Fi connection. But users will need to find a Wi-Fi hotspot, which, while widespread, is not as ubiquitous as cell phone networks.

“Speed rates [of Edge] will vary widely,” said Joe Nordgaard, managing director of wireless consulting firm Spectral Advantage, in an interview. “The user experience will vary accordingly.”

Nordgaard said users in congested areas later in the day when usage grows would find that their iPhone network connections will degrade. “It will help to be under a tower,” said Nordgaard. “Particularly for Edge it will help to be near a tower.”

“Wi-Fi will be the saving grace for the iPhone,” he added.

Even influential reviewer Walter S. Mossberg of the Wall Street Journal, who tends to adore all things Apple, called Edge a “pokey network” in his review of the iPhone this week”.

The data speed, Oh the speed! The carriers cannot hide their network data service’s mediocre performance anymore. Could they do better? Not so easy. Would 3G, WiMax or any other “magic” solution help? Some of the writers quoted above imply that it may. Well…sorry guys, it will not!

We have seen statements regarding the superior performance of “real 3G” networks. I am sorry to splash cold water on those hopes. While it is true that EDGE radio has slower speed than EVDO/HSDPA (great names) promoted by the carriers as the real stuff, it takes only a small fraction of the spectrum bandwidth. Do we remember what is the carriers’ most expansive asset? Why is it so important? The carriers’ available spectrum determines more than anything else the data speed they can deliver to the mobile subscribers. It really matters very little what technology they are using since the amount of spectrum is rather limited so the amount of service is actually preset.

Let us give you some perspective on data speed: A cellular base station can deliver about 2Mb/s per square mile per 1Mhz spectrum slice (this is very rough and also optimistic). Normally this base station serves about 60 users (voice) and that requires about 600Kb/s. Data service can use better transmission management (they call it transmission scheduling) so you can do better (~2MB/S). This limit has very little to do with transmission technology selected: 3G, WiMax, “4G” and any other magic names. Please refer to the article “A system has to know its limits” in this blog.

If the number of active data users is similar to number of voice users than the average data speed per user will be ~34Kb/s. Not much better than an old PSTN connection. So to get a speed of, say, 100Kb/s the number of active users per cell must be reduced to only 10 per cell-site. Now it is well known that there is an approximately fixed ratio of active subscribers (actually moving data) to the total served population (“pops”). If number of active subscribers is reduced, than the size of the potentially served population must be decreased as well. If we take 10 active subscribers instead of 60, the size of the served population must shrink as well by the same factor. If the carrier wants to keep the income per subscriber unchanged, it means the income per cell site is reduced by factor of 6. Saying this in another way, reducing the number of active users per cell by 6X means that we must add 6X more cell sites to accommodate all the users. So if we do not change what we charge is user, the income per cell site is reduced by factor of 6.

However, when more spectrum is used, achieving same coverage requires more transmission power; for example, if ATT switched from EDGE to HSDPA, serving the same area with same performance will require ~25X peak transmission power. This could be very challenging for the base station and a huge challenge to the iPhone mobile terminal.

If the carriers want to maintain their revenue at the current level they will need to increase the charge per subscriber by the same factor of 6. Does anybody believe that this is possible? Say typical subscriber pays about $50 per month for voice service, would the subscriber pay $300 to get 100KB/s service? Some argue that 100KB/s is not enough anyway, so the story may get even more interesting.

The bottom line is that the economics of providing high-speed access to a large number of cellular subscribers, whether by 3G or WiMax, simply does not compute.

Advertisement