Seen at DAC: are iPhone and Android showing the direction of tomorrow's handset architectures?

Apple’s iPhone and Google’s Android represent very different concepts of a cell phone—at least on the surface. But if we put the two of them side by side, could we conclude something about the future of smart phone design? That’s what two area experts, Portelligent president and chief wrench David Carey and ARM mobile solutions manager James Bruce, explored in a two-party panel in the DAC Pavilion last Wednesday.

The two are well-placed to have a view on the matter. Carey’s strip-down report on the iPhone was among the first—probably the very first—to identify the ICs used by the obsessively secretive Apple. And Bruce, reaching out from ARM to the handset, software, and carrier communities, has fingers on each of these key pulses.

The two speakers opened the conversation by pointing out that the iPhone and the Android represent very different concepts. Carey described the iPhone as "a triumph of co-design, created from the outside in." In his view, the iPhone is the result of Apple envisioning a user interface, and then systematically building a surface appearance, an enclosure, electronics and software to fulfill that vision. He emphasized how pervasive this vision was in the design. "The sequence of operations that Apple envisioned even determined the placement of sensors in the mechanical assembly and the physical shape of the package—for instance, to optimize microphone sensitivity," Carey said.

Carey proceeded through a component-by-component analysis of the major parts in the phone, noting that Apple—characteristically—showed no allegiance to any one IC vendor, but picked and chose as they wished. And the company opted for a surprisingly low level of integration: separate radio chips on a different board from the Samsung application processor, a separate Wolfson audio chip, and even a discrete Broadcom touch-screen controller IC. In order to stay within their chosen form factor, the company resorted to stacked packaging in some cases. Throughout, Carey said, you could see Apple designers working to make the best compromises of features, performance, form-factor, and cost, but all within the framework of that user interface vision.

Bruce used two examples, both handset designs from HTC, to show just how different the Android concept is. "Where the iPhone is a very fine example of a closed hardware/software platform, Android is entirely different: it is a software platform designed to be ported to a range of hardware platforms," Bruce explained. "Android includes a Linux kernel and an applications stack executing on a virtual machine." Thus there is no such thing as an Android handset. There are handsets—and there will be many more—that run the Android software platform.

The two HTC models Bruce compared were the Touch P3450, an OMAP-based handset with a 2.5G radio and a bill-of-materials cost under $90, and the TyTN II, a full-on smart phone with high data bandwidth at the opposite end of the feature and performance spectra. His point was that Android didn’t specify a product design or a hardware architecture, beyond the minimum requirement of a 200 MHz ARM 9 CPU for applications processing. Handset designers can adapt Android to any of a huge range of feature sets and cost points, while preserving the openness of the platform.

And that flexibility led to the one major difference in architecture between the iPhone and, for example, the TyTN II, Bruce said. As Apple moves down the cost curve, it can follow an integration strategy that produces an SoC tailored to the use profiles it is seeing on the iPhone. An Android-based platform must remain open to third-party applications.

Bruce went on to emphasize the importance he sees in the open development environment. "We’ve seen a shift in focus from thinness to functionality in the high-end market," he said. "Now to succeed, you have to beat the quality of fixed-function devices like MP-3 players and digital still cameras, not just integrate a lot of mediocre functions. This makes the open development environment a key issue. It’s the users who will drive the applications, not the system operators."

Bruce said display resolution was a case in point. He expects to see higher-resolution displays on handsets. He said we will also see the first high-res projection handsets this year, as well as further experiments, like Apple’s gesture-based dynamics, on making the small display interactive enough to present the user with a large data plane.

Asked if the system operators didn’t still have a veto power over handset features, Bruce replied that in the US they did, because a particular handset was tied to an air interface, and that meant it was tied to an operator. But in Europe, he pointed out, where everything is GSM, users can pick their handset and its features independently of their network operator, so it is in the interest of the system operators to support handset features.

Another audience member asked what the next big feature would be. Bravely wading in, both Bruce and Carey responded. Carey suggested that electronic-wallet features, similar to those already widely deployed in Japan, might come to the US, if the infrastructure investment can be made. Also, he is optimistic on location-based services, such as a tie-in to Google Earth. Bruce agreed, saying that electronic payment via handset will happen even in the US. And he suggested the next big thing might well be a new, as yet unspecified, service that fuses a combination of handset features in a way no one has yet tried.

So what can we learn from the iPhone and Android? One take-away is that the market has evolved past the point where you can be successful with good looks, or with a powerful platform that offers few services. The next successful phones will use their hardware and software platforms—whether proprietary or open—to deliver a complete user experience, fusing hardware and software functions into providing information the user wants with a minimum of user input. In this light, the two brands represent not two different architectures, but two different philosophies—the philosophy of the classical design house, imagining what you don’t yet know you will love, and the philosophy of the open system, allowing the market to create its own product on the fly.