re: MCI on Mobile Java and BREW
Many of ... short-comings in J2ME should, however, be remedied with the release of the MIDP NG specification, which should be completed by April 2002. Enhanced security features, essential when it comes to purchasing downloadable applications, and more formalised download provisioning are key improvements.
>> Talking The Same Language
Trefor Moss
Mobile Communications International
Issue 89, 01 March 2002
The vast majority of operators have come to consider java as a linchpin of their future plans for leveraging revenue from data applications. But how exactly did something that was never designed for wireless purposes acquire such rare status?
Big ideas are like gold dust to operators right now. The economic uncertainty under which the industry continues to labour has left 3G licence expenditure looking doubly injudicious and, despite continuing delays, those new networks are now being built - and paid for handsomely. Worst of all, these impressively expensive assets won't be worth twopence unless they are used to deliver services that will tempt customers to part with more cash, more regularly.
How fortunate, then, that there are some big ideas out there. Some of them, such as multimedia messaging and location-based services, remain largely conceptual and untested. Java cannot, in a way, be compared with these, being as it is a tool for building services rather than a service in its own right. But Java is commonly spoken of in the same breath because it has come to occupy a prominent place in the strategies of most forward-looking operators - and not only for 3G but for the generation of data revenues in the more immediate future. And if anything, Java is the most valid proposition of them all because it is already here.
Not that Java was ever intended for the wireless industry limelight, at least not at first. Originally a child of the 90s PC and internet explosion, Java was unveiled by Sun Microsystems in 1995 as a more user-friendly, platform-independent alternative to the C++ programming language. However, the 'write once, run anywhere' slogan with which Sun marketed Java attracted the attention of wireless players searching for ways to achieve their collective aspiration of a profitable mobile data market.
So it was in fact the wireless community, rather than Sun itself, which took the initiative to mobilise Java. "Companies like Motorola, Nokia, Siemens actually came to us," recalls Eric Chu, Sun's software systems group marketing manager, "and said that from their perspective Java was exactly what they'd been looking for." This interest ensured that when the Java 2 platform was released at the end of 1998, one of its three editions, the Java 2 Micro Edition (J2ME), was tailored for mobile devices. (It's worth pointing out at this stage that Java is both a language and a platform. So while there is only one Java language, there are three variants of the Java 2 platform each catering for the specific demands of PCs, network servers and smaller, mobile terminals.) There is a hitch, though. The 'write once, run anywhere' concept only applies up to a certain point where wireless applications are concerned.
The difficulty is that, unlike PCs, mobile devices come in all sorts of shapes and sizes, and disparities in screen size, colour or monochrome displays and sound capability are, if anything, set to increase. In addition, handset vendors are always inclined to take the existing standards and then expand on them in a proprietary way by adding in-house improvements so that they can differentiate their products. Developers will, therefore, need to make adjustments to their programmes so that the optimal experience can be delivered to each particular device. This should not prove a major problem, and the basic sentiment of write once, run anywhere ought to hold true. However, this will be an added headache as Java developers, used to writing applications for powerful desktop machines, learn how to adapt the Java language to the restrictive mobile environment.
But just how prominent a place has Java come to occupy in the mobile data proposition? "I think it's kind of at the centre of it all," argues Rich Green, Sun's VP and general manager for Java and XML software. Java's stock with operators has not grown simply because it paves the way for developers to produce compelling applications. Users are also given far greater scope for personalising their handsets by downloading over the air the applications that they want.
The will certainly exists to make 2002 the year in which wireless Java begins to make its presence felt on a global scale. "By the end of this year, half of all data-capable handsets will be Java-enabled," reveals Chu. "Nokia has said that they will ship 50 million Java-enabled handsets in 2002 and they're on target for that." ARC Group, meanwhile, predicts that there will be 110 million users of Java handsets by the end of 2002, rising to 421 million by year-end 2003. European operators are also well-advanced in their preparations for Java provision. Bruce Jackson, CTO of UK-based technology firm Elata, expects rapid adoption this year. "I'd be surprised if you could find a major operator in Europe by the end of 2002 who isn't trying to make money from selling and distributing Java applications," he predicts.
Telefonica Moviles added credence to these forecasts in February by announcing that it had licensed the Mobile Application Solution from US software firm 4thpass. The MAS system enables operators to manage and distribute Java applications over the air, and it will enable Telefonica, according to 4thpass, to boast Europe's first major mobile Java launch. 4thpass's marketing and communications director Javed Chaudry doubts that it will take Telefonica much time to roll out Java services, and the launch will take place "with five or six Java devices available. We expect that by Christmas time 15 European operators will have Java services." Chaudry adds that by the end of 2002, those operators will each have up to one million customers with Java-enabled handsets.
Java smartcards are also set to make their mark this year, according to Frederic Vasnier, Gemplus' VP for marketing. Gemplus, along with its collaborators in the SIM Alliance have ensured that "interoperability is now a reality" for Java SIM cards and they are expected to filter into the market this year, especially in GPRS terminals. Southern European operators have shown the most enthusiasm so far, says Vasnier.
A handset does not require a Java SIM card in order to be considered Java-enabled, for which the criteria are essentially the inclusion of flash memory and a virtual machine on which to run applications. A Java SIM in particular, though, provides benefits for "the part of an application which is related to privacy and security", Vasnier suggests. Gaming applications can still run on the handset, but financial applications, which include some forms of entertainment application such as gambling, can best be run on the secure Java SIM.
As with most new technology propositions, though, the solid proof that backs up the Java case is to be found in the Far East rather than in Europe. NTT DoCoMo launched its i-appli service - an i-mode extension offering premium Java applications - in January 2001. Within six months of i-appli's debut, DoCoMo was reporting that users with access to the Java services were downloading twice as much data as ordinary i-mode users, and developer numbers swelled upon the release of the i-appli SDK. By the end of 2001, over ten million DoCoMo subscribers owned Java-enabled handsets. South Korean operator LG Telecom also reported a significant hike in ARPU from users of its Java applications. Although at the end of 2001 only some five per cent of LGT's subscriber base was actively using Java services, the ARPU among that group was 76 per cent higher than that returned by other customers.
There were lessons for Europe to learn from the eastern experience, of course. The first concern was that Java handsets began to earn themselves a bad name for containing software glitches, with Japanese operators suffering a spate of handset failures last year. DoCoMo, for example, was forced into a mass recall of 420,000 Java-enabled Sony handsets in June because of technical problems, the inference being that the newcomer Java was to blame for any attendant gremlins. However, Sun's Eric Chu rejects the idea that Java was the cause of any of the recalls. "The fact that Java is the highest profile software stack in the handset is a marketing success for us," he observes, "but it leads to the misconception that if there's a problem with any of the software then it has to do with Java."
Then there was the question of standardisation. DoCoMo was so far ahead of the game as far as Java was concerned that it was ready to launch Java services before the Mobile Information Device Profile (MIDP), the initial wireless Java standard, had been specified. It therefore ended up using its own version, dubbed DoJa, for the i-appli service. This was all very well until KDDI and J-Phone began launching Java services too, at which point the 'write once, run anywhere' concept ceased to apply because DoCoMo's rendition of Java was incompatible with theirs.
But Japanese operators have at least demonstrated to their European counterparts that customers are likely to respond enthusiastically to Java applications, suggests Bruce Jackson. "They've all decided that Java is the right technology to use," he comments, "but they've gone ahead with proprietary handsets and implemented it differently." He believes that it has been worth Europe's while waiting for standardisation in order to avoid the incompatibility issues that Japan still has to face up to. However, DoCoMo can hardly be said to have suffered as a result of going it alone, quite simply because it has been making more money out of wireless Java than anybody else, with the likely exception of Sun. Having said that, DoCoMo has expressed a desire to come into line, and FOMA handsets have already implemented MIDP.
The US has also beaten Europe to the Java punch thanks to Nextel and Motorola's jointly branded iDEN service. iDEN may not have stolen the headlines in the way i-mode and i-appli have done - indeed, Jackson dismisses claims of iDEN's success as "more market bluster than reality." However, Nextel's director of wireless data services, Bob Ewald, defends iDEN as a pioneering product. "We now have over a million Java-enabled handsets in our customer base," he points out. Motorola has produced four Java handsets for Nextel, three of which have identical screen sizes. "We now have around 60 Java applications that are available for download," explains Ewald. He defends the argument that Java does indeed make developers' lives far easier and that "the 'write once, run anywhere' philosophy holds true with that exception of the display." Nextel has also followed DoCoMo's example whereby developers recoup the majority of the revenue - 80 per cent in the iDEN model - and developers are also given the power to set the pricing level for their applications, although Nextel has prescribed a minimum download charge of $4. Ewald's essential nugget of advice for operators looking to roll out Java services is that they should ensure that their packet-based networks are in place first.
The nearest the European market has come to a taste of this has come via mobile games service provider In-Fusio. Over-the-air downloading will, of course, be one of the main benefits brought by the large-scale Java launches anticipated later this year, but In-Fusio's ExEn game downloading engine has been enabling French and German customers of the Vodafone and Orange groups (provided they have the right handsets) to download games over the air since July 2001.
However, Java has still been at the heart of In-Fusio's offering. "We all realise that while Java has the basis to be the leading technology for entertainment-based applications on mobile phones," explains Giles Corbett, In-Fusio's managing director, "there's a whole bunch of things that are still missing." The games being downloaded are all written in Java, but the ExEn engine, which has been incorporated into 11 handsets built by Philips, Siemens, Trium and Sagem, includes proprietary improvements on the basic Java model that In-Fusio felt were necessary.
Many of what In-Fusio regarded as short-comings in J2ME should, however, be remedied with the release of the MIDP NG specification, which should be completed by April 2002. Enhanced security features, essential when it comes to purchasing downloadable applications, and more formalised download provisioning are key improvements. The present version of MIDP does not yet define a standard for OTA downloading. ExEn was once considered to pose a potential threat to the Java platform, at least on the gaming front, but Corbett now expects that "by the end of 2002 ExEn and future versions of MIDP will converge." To this end, In-Fusio announced in February that it was partnering with Sun and the Java Community Process in order to create a J2ME-based wireless gaming standard.
So if ExEn presents no risk to Java's progress, how do other technologies check off on the list of complements strike competitors? First up, of course, there's WAP. It appears that WAP and Java will co-exist because they don't really perform the same function. "WAP browsers are more designed to allow you to view and find information," says Eric Chu, "not necessarily for you to interact with information." For a rich, interactive experience, then, go with Java; for text-based information stick with WAP.
If Java does have a rival, then perhaps Qualcomm's BREW (Binary Runtime Environment for Wireless) platform is it. Apart from emulating the beverage-related nomenclature, BREW is billed by Qualcomm as "an open, standard applications execution platform for wireless devices." Doesn't that sound strangely familiar? Qualcomm thinks not, insisting that BREW is in no way an attempt to encroach upon Java's territory. "We do feel like we're complementary to Java both on the environment side and the reach side for application developers," explains Jason Kenagy, Qualcomm's director of product management for BREW.
A major distinction between BREW and J2ME is that BREW primarily targets C/ C++ developers. However, the platform also supports Java applications which, Kenagy observes, is good news for Java developers because otherwise they would have to write for "lots of different Java platforms." When exactly the one standardised Java platform fulfilling the 'write once, run anywhere' promise will arrive "remains to be seen", in his opinion.
The logic of encouraging developers to ignore Java in favour of C++ has been called into question, though. 4thpass's Javed Chaudry suggests that "after years of evolution in the PC industry very few application developers are focusing on C/C++ any more." There are conversely around two million Java programmers out there. Kenagy, though, is adamant that "C and C++ applications run [up to] seven times faster than Java apps." Speed issues will, of course, be resolved as handset processors become more powerful. Nonetheless, operators are beginning to look at BREW as a tempting proposition.
In February, Verizon became the latest operator to announce that it would be rolling out BREW services, following on from 15 other operators, including such big-hitters as KDDI, KT Freetel, Telesp Celular and China Unicom, which have so far said they will support BREW. In other words, a J2ME versus BREW split along the old GSM versus CDMA lines is not unlikely.
Still more difficult to call is the potential impact of Microsoft's .NET, the XML services platform. .NET is the unifying influence across all of Microsoft's business areas because, in much the same way as the J2ME, it gives developers common tools with which to write applications for any Microsoft-powered device, including smartphones. A Java virtual machine can be run on top of .NET but, like Qualcomm, Microsoft is leaning towards development in C++. Microsoft's wireless prospects remain entirely debatable, but if the Redmond giant does gain a significant foothold then .NET could really threaten the Java platform because Microsoft wields such an influential developer community.
In its favour, though, Java could have a head start of several years over .NET, at least in the wireless business. And particularly crucial is the momentum that Java has built up from the operators' point of view, to the point where it underpins their all-important transition from pure voice businesses to data service providers. Operators are certainly going to have to come up with some fairly stunning ideas in order to make the 3G gamble pay off. But Java doesn't look like a bad place for them to start.
- Eric - |
