Down the Slippery Slope
By Justin Ried, Mon Mar 19 00:00:00 GMT 2001

With projected 3G rollouts getting closer and closer, much of the industry seems to be getting nervous about delivering on time. In fact, a few major players just announced major delays. What?s got them worried, and where's my killer app? Updated.


On Tuesday last week, Japan Telecom (the company behind J-Phone) said it would push back deployment of its third-generation networks by a full year - to late 2002. Qualcomm CEO Irwin Jacobs told the Financial Times just recently that many 3G services won't be viable until 2005 - a full two years behind the company's previous projections. He went on to say that their UMTS networks, currently in testing, are still facing some "serious technical hurdles."

Michel Rahier, head of Alcatel's mobile unit, echoed that statement at the 3GSM Congress in Cannes last month by saying "The roll out of 3G will take place three to five years from now; Whereas last October we were saying two to three years. We're now looking at end-2003, early 2004." He went on to say "UMTS terminals will take over in 2007, not 2005. The pace is going to be a bit slower than we thought a few months ago, which means the GPRS lifecycle will be longer. The availability of terminals and services are key and this takes time."

In fact, NTT DoCoMo will be launching Japan's first third-generation services on May 30th, but there will only be two compatible handset models available when it goes live.

2004, 2005…2007?


So what's the hold up? For one, the network technology is still very immature. The CDMA-2000 networks Qualcomm's Jacobs mentioned are hampered by unacceptable handover times - requiring up to 10 seconds to complete the switch. And they don't appear to be having much luck when they decrease the data rate, either. It's not the high-speed flow that makes the chain easy to break, but the protocol's handover function that needs improvement.

The industry has thus far been focused on developing ways to squeeze as much data as possible into the narrowest bands of spectrum. Things like handovers are considered second, and the industry is learning the cold reality: What brings the most speed doesn't necessarily deliver the best voice services.

Actually, the industry as a whole still seems confused as to just what UMTS is: CDMA-2000 or W-CDMA. Qualcomm would certainly prefer the latter since they'd reap more royalties from it, but most of the industry seems to be banking on the alternative W-CDMA.

The platforms


Once the network is defined and operational, which handset platform will be delivering the services remains another unknown.

Microsoft is changing their dream of "a computer on every desktop" to "a computer in every hand." But changing the corresponding business model has proven a bit difficult. In fact, none of the top 5 mobile phone manufacturers (comprising over 60% of the market) have committed to using their new platform, code-named Stinger.

Symbian's got a good candidate in EPOC, which has been endorsed by the top three mobile phone manufacturers. It's a robust little OS with lots of third-party applications already written for it. But while the platform has the support, products using it have been slow to appear. In fact, the only one available right now is Ericsson's R380. The Nokia 9210 Communicator, the first to use EPOC version 6, is due out before the summer.

Version 4.0 of Palm's somewhat antiquated OS has just recently jumped out of the gate. But wireless elements are still tacked on, and telephony is still supplied by third-party developers. Version 5 of the OS sounds promising on paper: 32-bit, multi-tasking, and wireless at its core. But with no release date yet available, it may be too little too late.

The Java/Java2 Micro Edition runtime environment may simplify things with its "compile-once and run on anything" approach, but devices compatible with Java are behind schedule. In fact, several Java-enabled i-mode phones were released earlier this year, but they were re-called by the manufacturers shortly thereafter because of bugs and performance issues.

Today there isn't a single platform that operators and application developers can target when developing next-generation services. Software companies still have to hedge their bets when considering what platform to develop for, and that means their hands are tied.

The applications


Here's the factor that should be driving the devices to market as soon as possible: Killer applications. But because of reasons outlined above, there is no single killer application that has consumers chomping at the bit. Their collective desire still resides with "faster, better, cooler" but not with any single compelling service. Consequently, their attitude is more of "that'll be nice, when it arrives" instead of "gotta have it, need it now!"

We've heard quite a bit about location-enhanced services, mobile multimedia, broadband Internet access and dynamic, multiplayer games - but where are they? Who's going to be offering them, and perhaps more importantly, how much do they cost? No one's sure just yet. Because of constantly slipping launch dates, operators are hesitant to advertise these. They got their fingers collectively burned by prematurely launching over-hyped WAP services last year, and it seems they've learned their lesson.

There’s always a “but…”


However, remember those CDMA-2000 networks I just mentioned, the ones having handover problems? Well, surprisingly enough, according to an Justin Ried .