There's no doubt: The cloud makes mobile computing more useful. But that truism can quickly lead you to a bad place -- namely, that mobile devices should just be portals to cloud services. I was reminded of that with Amazon.com's announcement last week of its AppStream service, which lets game developers run the compute-intensive rendering for their games on Amazon's cloud service and deliver the results to mobile game players over the Internet. The pitch is that basic smartphones without high-end GPUs can still play graphics-rich games.
Yes, this is more of the interminable native-versus-cloud argument that's been raging for several years, though in a different context. Here, the pitch isn't primarily about cross-platform deployment but about overcoming hardware limitations. Still, it's a false argument.
[ Apps with a deathwish: How to ruin a perfectly good mobile app. | Heed these 10 expert tips for mobile app design. | Keep up on key mobile developments and insights via Twitter and with the Mobilize newsletter. ]
Game playing has quickly migrated to mobile devices, led initially by the iPod Touch and now on bigger-screen devices like the iPad Mini. Apple's devices made in the last few years have sufficient processing power to handle most of today's graphics-rich games, as do most Android devices from name brands like Samsung, Motorola, and LG. But Amazon says users of less-capable mobile devices can run such games too, if assisted by its AppStream service, which will do the heavy lifting.
It makes sense for Amazon to say this: It makes a lot of money providing cloud services, and no doubt it would like to make game developers dependent on its services, as well as ultimately offer subscription-based gaming over that service, a Holy Grail for the gaming industry that wants to smooth the financial ups and downs of relying on hits.
But the cloud is a bad option here. One reason is that you need a fast broadband connection to do Web gaming, so play becomes limited to people's living rooms and offices. The mobile networks, despite what the carriers advertise, don't have the capacity for graphics-intensive gaming any more than they do for TV delivery. Heck, I often can't even get Twitter to download new tweets on my train during commute hours.
Broadband access costs real money, so if the pitch is to widen the pool of players to include those who can't afford higher-end mobile devices -- well, chances are those people don't have fast broadband at home, either. With carriers implementing data caps on home broadband, moving users to online game playing will only increase their costs, especially if they're also using other services like Netflix and Hulu.
A related reason is that you need a consistent connection. You probably get that at home and at the office, but few other places. The Wi-Fi at hotels and cafés is often slow and intermittent, for example. And the cellular networks often have weak signals or are overloaded. Caching can overcome these issues only so much.
Graphics processing is only part of the puzzle, as Microsoft has found in its experiments around cloud delivery of Xbox games.
The truth is that native processing simply makes more sense for game-playing than cloud augmentation, much less delivery -- as it does for so many other things. Companies like Amazon and Google of course are betting on cloud-based services, so they want to steer both developers and users to them, but that bet is charitably a long-term one.
The cloud is a great place to get data from in burst-style usage (à la email and the Web), and it's a great supplement to mobile devices' innate capabilities (à la cloud storage). But it shouldn't be used as a critical component of your mobile computing engine.
This article, "Mobile doesn't need the cloud as its engine," was originally published at InfoWorld.com. Read more of Galen Gruman's Mobile Edge blog and follow the latest developments in mobile technology at InfoWorld.com. Follow Galen's mobile musings on Twitter at MobileGalen. For the latest business technology news, follow InfoWorld.com on Twitter.