March 2009 Archives

"Figures Lie and Liars Figure" is the common statement about charts, so I am look at the notes about smartphone OS with a grain of salt.

Erick Schonfeld at TechCrunch posted these numbers based on Ad Mob statistics which is a skewed sample to begin with, but at least has some validity for the category of those who build apps looking for ad revenue.

Let's talk about the numbers.

Apple iPhone is now 50% of the smart phone traffic in the US. This represents tremendous growth, because it's an analysis of the drop of water of the small pond of Internet traffic. Because of the slice by device analysis the OS of Symbian, RIM, Palm and Blackberry are the only contenders. And as would be expected starting from zero Android is the only other one with a gain. While Palm, longest in the tooth, is the lowest man on this totem pole.

There are some good points in the details that are worth considering, particularly the fact that Storm is lagging behind Bold in the RIM world. It maybe the navigation issues are still problematic.

So what don't I like about this analysis. Lots of things.

First of all, I would want to analyze is how many of these devices are using their dual mode ability. I think WiFi makes the difference for a lot of these devices and therefore the dongles to the PC are still the winner for mobile Internet use.

Secondly, the apps of iPhone need to be further explored. Is the browsing the traffic, or is it application specific based traffic. Or is this only based on the ad push that we are seeing this data?

Finally, can we talk about this in terms of overall mobility and not just this one category of devices.. Most mobile data traffic is from PCs, Using these figures, I could say mobile data traffic from PC's is on the decline which is just not the case.

When ever I talk to application developers about communications solutions the discussion about the web versus the phone developers is emphasized.  Millions of Web developers compared to a few thousands of phone specific developers.   The point being that of course the innovation is with the web.
If you look under the hood the chips inside for the web are the same on your PC. If the developers start on a PC Chip, it's kind of logical to expect that devices built on PC Chips have a head start.

It also means a shift in the way to look at innovation.  For example, if you have a new codec developed in software that you built on a pc Chip and the cost to go to market, start a trial are incremental if the devices are also pc chips.

Ultimately, innovation then is not about the store, it's just about the Internet

The New York Times this week had a great article about the computer chip manufacturers coming into the telecom industry.  The article had lots of good references but I think it missed the major part of the story.  Phones, even smart phones, are carrier specific controlled solutions.  Even in the GSM world the carriers have a tendency to put a customization on their phones.
In the computing industry, the consumer actually has less selection than in the phone industry.  However in the phone industry you end up obligated for a while.
The real story with the iPhone is the standardization of single device that is resold by the service providers without the customization.  This is more important as we evolve to the wireless broadband. 
As the ability to provide applications in the cloud of networks, the device that internetworks is going to be the story. 
So why call this a smartphone? 
Candidly, on the phone side, David Isenberg's dumb network is being matched like for like.  The features of so called smart phones are not the phone side.
Mobile Internet Devices is a term that does not seem to have captured the public vision, but netbooks is a term that has caught on.
To my consumer (untrained eye), a netbook is a retro laptop that gives me the access I want.  The price point is under many of the smartphones and if it had a data plan with it, I would be happy.
Maybe we should be looking to blend the terms.  Netphones, perhaps.
I like this as a generic term since it does not give me the false expectation that the phone is smart.  It also speaks to where we should expect the innovation to occur the blending of the net with the phone.
It also helps put the emphasis on what a network operator should be concentrating on enabling the broadband and blending the web.
It also changes the expectation of the device, since its no longer about the subtlety of a carrier's specifications, but a manufacturer's standardization.
This is also changes the game for manufacturing.  A well known cell phone manufacturer makes over 3,000 types of cellphones. 
I don't believe this is netphone future.  All the peripheral casings you see at kiosk's may have a life, but consumers choices will be based more on their own taste, then the carrier's projection of their choice.
And that's what will be smart about phones in the future.

One of the next major standard strategies for 3GPP is LTE Advanced. In Release 8 we have SC. Today I am at our friend Dr. Hyung Myungs's Workshop at NYU Poly about Single Carrier - Frequency Division Muliple Access [SC-FDMA]. 

I still have to ask? What is the mission?

The Vision of the 4G Broadband Wireless Era is being discussed around me today. Its a lot of math and its focused on the PAPR problem. Listening I am impressed with the work, but still not sure the premise is right. What devices are we going to be using in the future? In the presentations, I see the goal as managing the issues of frequency. Slicing the channel with OFDM in a distributed mode, makes the most sense to me as I listen and grapple with the calculus.

However, It feels like we are still coping with a world that may not exist. In IP particularly with UDP, losing a packet is no big deal, and in TCP the ability to resend saves the day.

So when does the problem show up?

The answer back is when we have Interference in the signals, but precoding is being done and the signals are being managed at several layers.

I think WiMAX maybe right to take a pass on implementing this standard. If the device of the future is not really a phone the processing on the edge maybe able to over come a lot of these problems. For many devices and applications SC role has somewhat limited value. It maybe that Video is the most likely benefactor of this next generation.

Of course it may be that the amount of devices become so numerous, the interference makes SC essential.

As is so often said this would be a nice problem to have.