Hard Cell

Japan's cellphone giants are about to conquer the planet, right?
Wrong.
We're not talkin' boomboxes and Walkmans here, and software integration and internationalization are serious barriers.

In the first half of this year, even as Japan mobile Internet usage was going through the roof, a number of embarrassing and expensive cellphone recalls involving NTT DoCoMo's i-mode service occurred -- and we're not just talking about an LED that won't light on one side, or a recalcitrant On/Off switch. The recalls involved catastrophic software failures with the new Java phones, including address books getting wiped out, saved email messages (all of them) being erased, and downloaded-and-paid-for Java programs being reduced to digital dust. Think of these malfunctions as the keitai equivalent of Windows' infamous Blue Screen of Death, except that there's no "reboot" button on a cellphone. Subscribers were furious, and on each occurrence, the network operator and handset makers rushed to offer replacements and a heartfelt gomennasai. Press releases put a soothing spin on the situation, using language like "rare instances" and "free exchange."

But few have asked what these software failures imply for the future of mobile computing, and in particular whether these episodes call into question Japanese handset makers' ability to innovate and integrate ever higher quality mobile hardware with ever more sophisticated software. And the challenge of keitai construction becomes even more difficult if you consider that the makers are under intense pressure to export their top-of-the-line models overseas, something they haven't been able to do for a complex mix of reasons related to technology, economics, culture, and a self-absorbed fixation on the domestic market. Internationalization of the pocket rockets seen so far only in Japan would be tough enough even if all the onboard software did work as intended.

It would be tempting to write off DoCoMo's 2.5G i-mode Java glitches, as some pundits have, as mere teething problems at the dawn of the wireless Internet era; inconvenient, perhaps, but isolated cases that the engineering powerhouses behind Japan's cellular industry will soon have licked. But on April 26, NTT DoCoMo, one of Japan's (and by extension, the world's) most capable and experienced cellular operators, shook the global wireless industry to the core by announcing that the full rollout of its much-heralded 3G network would be delayed until October while it initiated a phased trial period to help test the network.

Speculation attributed the cause to software problems with the new handsets and with the unproven W-CDMA (Wideband Code Division Multiple Access) cellular base stations. In a tacit admission of this reality, the press release said, "NTT DoCoMo believes that further testing is needed in order to guarantee an even higher level of system stability." DoCoMo was also hobbled by the fact that only two of the 11 vendors that had committed to provide 3G handsets were ready in time for a May launch. "It's kind of anticlimactic -- we knew that not all of the handset makers were ready," says telecoms analyst Mark Berman, at Credit Suisse First Boston. "We figured they'd end up launching [only] on a limited scope. DoCoMo is finally backing away and admitting, 'Hey -- it doesn't make sense to do this.' " Suddenly, the supposedly isolated 2.5G i-mode Java glitches appear to be part of a larger pattern involving problems with the handset and network software on the 3G system as well.

Many observers are now asking how 3G can succeed when even the (relatively) low-tech Java software on the present i-mode handsets is buggy. And it isn't only DoCoMo that is seeing problems with handset systems integration. In March, competing operator J-Phone announced the delay of its 3G network rollout (until June 2002), in part because "the terminal makers are not yet ready."

"We cannot eliminate the bugs [on cellphones] completely, said Koji Nishigaki, president of NEC, in a May newspaper interview, adding, "The handset itself is like a computer." In early May, British Telecom recalled handsets recently received from NEC for use with Manx Telecom's new 3G network. The cause? Software glitches.

It's clear that cellphones, with their ever more complex onboard software, are fast becoming one of the most sophisticated examples of personal high-tech engineering used in daily life, and that the Net -- or the browser phone -- is forcing Japanese electronics firms to finally confront software. It's no understatement to say that the software problems have caught the Japanese elec tronics makers, long venerated as wizards of hardware manufacturing and miniaturization, with their zubon down. Previously, they only had to deal with deeply embedded software, if that. To export a Walkman or a MiniDisc player to the States or to Europe, Japan's consumer electronics giants needed hardware but not software expertise.

Compounding the difficulties of the new paradigm is the fact that a cellphone is fundamentally different than a Walkman or a car stereo, devices that work much the same no matter where they are used. Making such stand-alone electronic gizmos requires a domestic-based hardware capability that can be fostered entirely in Japan, in Japanese, and without having to deal with messy overseas business entanglements. When it came time to ship some overseas, Japan's fabled shosha, the general trading companies with links into virtually every market of any worth on the planet, would traditionally handle the shipping, the customs clearance, the inland distribution, and finally remit the profits straight to corporate finance, thank you very much. Translating the user manual into English (or French, or any other language) was an administrative challenge, not an engineering one -- and there, too, Japanese electronics makers demonstrated that their expertise lies in hardware, not in translation, as anyone who has ever struggled to understand their VCR programming instructions can attest to.

But cellphones make use of publicly owned airwave spectrum, and their design, engineering, manufacture, and operation are all intimately tied up with national and international standards bodies, global telecommunications interests, local government regulations, and conflicting user cultural demands. The new, uniform third-generation W-CDMA network standard not withstanding, a cellphone made for sale in one part of the US is still a different beast -- in terms of chipset, hardware standards, software protocols and functionality, and onboard subscriber services -- than one made for sale elsewhere in North America, Europe, or Asia.

Now that the browser phone is turning the Net into an onboard consumer electronics feature -- and is clearly the hot new accessory -- Japan can no longer stick to hardware only. To succeed domestically and win lucrative slices of overseas markets, Japanese makers must confront software and, more hellishly, software integration.

SOFTWARE IS HARD
When it comes to software, the Japanese makers may finally have met a challenge they can't overcome, and industry watchers are worried about the makers' ability to deal with the challenge of onboard software and systems integration -- getting everything to work together consistently and smoothly. Before, all the software was embedded, and functionality was limited to driving the hardware (power the onboard cell network chipset, drive the speaker, or pass 8-bit data from one module to another), or displaying simple information like which phone number had just been dialed.

Now, like a PC, cellphones comprise multiple layers of software (see figure on following page), including embedded driver software, a micro operating system, an application interface layer to connect onboard applications to the OS, and the applications themselves, including the all-important microbrowser and its user interface, which is the only component of the software micro-hierarchy that the user usually sees. This software is the platform over which the network operators wish to provide ever more lucrative wireless information services -- the type that users demand and will pay for, including e-commerce, financial services, navigation, database access, and unified messaging.

"Three years ago, we only saw small applications on the keitai," says Access CTO Tomihisa Kamada (Access makes the dominant i-mode browser -- see "Microbrowser Battles"). "Usually, nothing more than an address book." Now, he explains, cellphone software includes Java (for security and to run downloaded applications like games or stock trading services), comprehensive mail applications, and other functionality, like navigation, multimedia, and Web browsing. "This is all new to the makers," he says, adding, "Now the big makers are all trying to hire more [software] engineers and transfer existing ones into their mobile divisions." Taro Yamanaka (not his real name), a technical marketing specialist at Motorola Japan, agrees. "Software is more complex than three or four years ago. The trigger was i-mode. Most Japanese makers are now struggling to increase the number of engineers assigned to mobile," he explains.

CELLPHONE SOFTWARE ENVIRONMENT

Note: Services like security now must be provided in all layers. Source: J@pan Inc research

Most of the new applications that subscribers demand must be implemented via software. "Mail, games, e-com, network downloads -- [whatever] the user needs," says Hiroshi Sakai, director and senior general manager of KDDI's Mobile Communications Engineering Division. And these demands spring from the most mundane of needs. "In Japan, people tend to read email while walking, which is dangerous," says Masaaki Yamamoto, senior manager at Openwave Japan. "So voice readback is a key function." It takes sophisticated software to achieve such functionality.

Also, keitai software functionality is tightly related to the hardware, similar to the situation at the very start of the PC revolution. Back then, hardware capabilities quickly exceeded software programmers' needs, and PCs have long since become far more powerful than what is required for typical uses (word processing, Web browsing, email, file download, et cetera). On the keitai, however, memory size, processor speed, and battery life are still fundamental constraints for both cost and physical packaging as well as software deign. "We always want more memory," says Access' Kamada, "but makers always want to reduce cost. Even if the handset is very functional, no one will buy it. Balance is key." Handsets have, in fact, grown slightly larger recently, to accommodate more chips, better displays and speakers, and bigger batteries.

A further challenge for software developers is that the systems integration must span software operating in several different physical locations connected by an air interface that presents latency and circuit-connection issues unseen on the Internet at large. "The integration between the network, the Web server, and the handset must be tight," says Motorola's Yamanaka.

But not all the demand for enhanced software integration is coming from the end-user. The handset makers themselves are struggling to reduce the cost to develop new models, deploy new services, and speed time to market. What limits them is their long-standing cottage-industry approach to device software, in which each new model of a particular product has unique new software developed for it, and the work is divided across a number of divisions.

This approach stems at least in part from a long tradition of domestic competition and engineering pride. "Lots of Japanese makers have amazing labs, but they produce proprietary technology," says Francis Charig, chairman of Tao Group, a maker of embedded operating systems that has operated in Japan for a number of years. Charig explains that makers are realizing the value of a unified look and feel across multiple product lines, and that the user interface, whether on the keitai, PDA, fax machine, or PC, should present the same brand image to the customer. He points to the success of Sony's Vaio series of PCs and PDAs, many of which comprise common software, common Memory Stick removable media, and the common color deep purple. "Customers only want one system to learn, not many," he says, while "the makers want a strategy around which they can build up their software."

Indeed, in a clear acknowledgement that a comprehensive approach to electronics manufacturing -- including software development -- is required, Sony announced in March that it would integrate its five existing network companies, organized along product categories, into seven "solution-oriented" network companies, on the assumption that "game and Internet/communication services will be deeply integrated with [the] electronics hardware business in the broadband era."

OVERSEAS EXPANSION: A HARD CELL
Despite their long-standing hardware expertise and moves to solve the software problems, it's still going to be tough for the Japanese makers to compete overseas. "The European and Japanese markets are completely different," says Sebastien Marteau, marketing director for Asia-Pacific at leading smartcard maker Gemplus. In Europe, he points out, all 2.5G and 3G phones must be smartcard-enabled, a requirement that doesn't exist in Japan. In the US, the FCC (Federal Communications Commission) has mandated that all cellphones be able to report their location via the network to emergency services so that car accident victims, for example, can be found even when the caller doesn't know where he or she is. These requirements don't exist in Japan.

The problem of packaging and integrating superior-quality components and software at a price consumers overseas will pay is a major challenge. Japanese makers have little experience making handsets for GSM (Global System for Mobile Communications) or other standards found outside of Japan. So far, they've concentrated most resources on creating models compatible with Japan's PDC (Personal Digital Cellular) network, which is only used in Japan.

It's also clear that the highly developed and complex system of cellphone subsidies in Japan has played a major role in getting superior quality devices -- especially color displays -- into consumers' hands (for a detailed explanation of how these subsidies work in Japan, read "Cellular Invasion Gear"). Whether consumers overseas will pay for machines similar to the ones now sold in Japan, absent any subsidization from the operator (illegal in some jurisdictions, frowned upon in others), is an open question.

The first full-color phone available in the US, the Sanyo SCP-5000 offered by Sprint PCS, admittedly is pretty cool, and appears similar in weight, design, and quality to several of the current models popular in Japan (it boasts 256 colors and a standby time of up to 120 hours). The problem is, it retails for a wallet-busting $499.99. A similar 256-color phone in Japan, the Panasonic P209is, with a standby time of 380 hours, debuted last summer at around ¥30,000 and can now be picked up for less than ¥10,000.

It seems the Japanese will be in for a rude shock when they try to market their top-of-the-line pocket rockets at full price to European and US keitai surfers absent the subsidies they've long been guaranteed back home. "Even though Japanese makers already sell in Europe, there are no color displays or 64-tone sound chips [on their European models]," adds Gemplus' Marteau.

Overseas keitai culture is also a challenge to Japan's makers. "It's really a matter of content," says Openwave's Yamamoto, adding, "No one can predict what people will pay for." He points to the usage of mail in Japan versus that in Europe. "In Japan, it's all email, while in Europe it's all SMS." (Short Message Service is a hugely popular text-based mail service similar to Internet email, but it's necessary to use a gateway if subscribers wish to send an SMS message from a cellphone to a PC.)

Culture, of course, refers to the software developers as well as the mobile users. Yamamoto mentions India, where English is widely spoken. "India's a software powerhouse. Japan isn't," he says. And there's at least some evidence to support the argument that Japanese keitai makers will be challenged by language. In April, Java developers creating games and other programs for DoCoMo's i-Appli Java-enabled phones found out that one model expected the downloaded Java application to report its date as "Apl" instead of "Apr" -- a simple but telling indication of just how far onboard keitai software has to go to meet international standards.

Japanese makers may also find it tough to create handset distribution channels and sales, service, and support operations in Europe and the US, unless, like Sony, such a system already is in place to support existing consumer electronics sales. Creating localized user manuals and sales material will be another cost. Partnerships with foreign technology companies could be key to easing the challenge. In April, Sony announced that it would join its cellphone terminal manufacturing efforts with those of Swedish telecommunications giant Ericsson. The deal will allow Ericsson to benefit from Sony's design and marketing expertise, as well as its experience with i-mode, while Sony will get the benefit of Ericsson's established worldwide distribution channels. (Last year, Ericsson had about 10 percent, and Sony less than 2 percent, of the worldwide handset market.)

In November last year, Toshiba signed a deal with Siemens covering development of 3G handsets, and Kyocera has already started marketing a phone in the States that is converged with a Palm OS-based PDA. Mitsubishi, Matsushita (Panasonic), and Sanyo are all reported to be in discussions with foreign makers for various types of cellphone cooperation.

The Future
In the future, wireless services implemented via software will only become more complicated, and the complications will extend across all portable wireless-enabled devices, not just cellphones. P2P networking may be huge, and may account for a large component of corporate applications. "P2P data transfer is significant," says Lou Pirrelli, VP for sales at Web storage firm i-drive.com. "P2P is not only device to device; it can be sharing regardless of the data's location." This clearly involves a sophisticated onboard environment -- precisely what the Japanese makers are having trouble creating.

Ultimately, it may be the advent of 3G -- currently causing the headaches for handset makers -- that enables the likes of Sony, NEC, Matsushita, and Fujitsu to really bring their software development firepower to bear on solving handset problems. Until now, Japan was the only country in the world using the PDC (Personal Digital Cellular) standard mandated by NTT. "As a result, Japanese makers couldn't participate [fully] in the European or US markets [which use the GSM and TDMA/CDMA standards, respectively]," says Kazuhiko Shirai, who worked in terminal manufacturing at Sony before joining PacketVideo.

Most 3G networks, however, will use common W-CDMA protocols, and it will be much easier to make handsets that can be sold into multiple markets. The onboard systems integration, therefore, will only have to be done once. Maybe the Japanese makers' time has come? "In 1G analog, Motorola dominated," explains Access' Kamada. "In 2G, it was Nokia and Ericsson. Now, with 3G, everyone uses the same network infrastructure, so I expect the Japanese makers to do well." Paul Rogers, marketing director for mobile networks at Logica, agrees: "With 3G, Japan rejoins the world, and the PDC/GSM wall is gone." Rogers goes so far as to speculate that Sony could become the strongest handset vendor in the world, displacing Nokia from that lucrative position.

And whatever the market, handset makers everywhere are constantly under pressure to deploy new features and services from the growing legions of third-party content, application, and service developers focusing on the wireless Internet. "We want to see on-the-fly file download, local storage of security information, better memory and processors, bigger screens, and voice recognition," says John Sims, CEO of Texas-based mobile financial solution developer 724 Solutions, referring to the capabilities he'd like to see in the short-term future.

No one's worried that the Japanese makers won't be able to innovate their way out of any hardware issues and provide the best integrated audio, graphic, video, and multimedia components that the ultimate end-user -- whether in Seattle, Sendai, or Stockholm -- is willing to pay for.

But it's not certain that they'll solve the software challenges. And if they can't, maybe cellphones of the future will come with a reboot button. At least you can rely on the Japanese to select a cuter crash color than Microsoft's stark, dark blue screen of death.