The Shape of iPads to Come
by Mark W. Hibben
Prediction Tempered by Experience
As the end of the year approaches, predictions as to the next generation of iPad already are starting to appear, along with leaked images of would-be iPad killers, such as a new Motorola 10 inch tablet based on Android “Honeycomb” OS (next in line after “Gingerbread”, which is coming out with the new Nexus S). So this seems an opportune time to go on the record with my prognostication for the next iPad. In the past, my predictions about Apple products have tended to underestimate the conservatism and risk aversion of Apple management, but after almost a year of closely analyzing Apple decision making, I think I’ve got them pretty well calibrated. Therefore my predictions are based on “best decisions” within the Apple corporate culture and mindset. By the time that the iPad 2 appears in stores next Spring, it may well emerge that what Apple management viewed as the best course of action and the actual best course have begun to diverge.
Even as far back as the early days of Mac development, Apple was less about technology development than about early adoption. Apple executives toured Xerox Palo Alto Research Center, saw their clunky wooden mouse and primitive experimental GUI, and adopted the technology ahead of the competition. The fact that Xerox was never able to capitalize on their own technology development (TD) in this area of GUI, shows the risks inherent in true technology development. TD is always a crap shoot, and even when something comes of it, the company may not know how to exploit it.
Apple doesn't really do TD for the most part. What Apple does, and does very well, is creative systems integration, the combining of hardware and software technologies to yield novel and often compelling consumer products. MP3 players predated the iPod, but Apple was able to combine the player, a laptop hard drive, and superior software in iTunes to make a winning product. No TD required. This approach has the advantage of keeping R&D costs low and increasing the probability of the research yielding a viable product. But it does limit the kind of innovation of which Apple is capable.
As Apple has grown larger and more mature, it has to some degree deviated from the path of systems integration/product development into some true TD, specifically in the capacitive touch sensor technology that forms the backbone of all of Apple's touch screen devices, as well as the track pads in Apple notebooks and the desktop Magic Track Pad. Here, the TD investment has paid off handsomely, but largely because Apple understood how to adopt and integrate the technology into their products.
The iPad showed the world just how creative the Apple systems integration approach could be, producing a truly unique product out of almost entirely “off the shelf” components. Once again, there was little real technology development. The main technology development was in creating the larger area touch screen interface. Almost everything else about the product was derivative of the iPhone, which in turn derived from Apple’s early tablet development efforts.
Even the A4 processor, which Steve Jobs hailed as an Apple development, turned out to be an exercise in early adoption and systems integration. The iPhone and iPad use systems-on-a-chip (SOC) processors that are very much equivalent to an entire computer motherboard of a few years ago. The SOCs combine central processing units (CPUs), memory interfaces and various peripheral interfaces such as USB as well as video graphics processors. What Apple had done in “creating” the A4 was apply their tried and true systems integration approach to the SOC level, combining component parts of the SOC licensed from various vendors, then giving the completed design to Samsung to manufacture. Even the CPU core was licensed from Samsung, which had jointly developed it with the American processor design house Intrinsity. Interestingly, Apple subsequently bought Intrinsity, sparking a rumor that Intrinsity had designed the A4 CPU core for Apple, but the design was completed before Apple bought the company. Samsung now uses the same CPU core design in an SOC of their own, which goes into many products competing directly with the iPhone and iPad, usually running the Google Android mobile operating system.
And Now the iPad 2
I’ve recapitulated some of the recent history of the iPad by way of leading up to my predictions for the iPad 2. I expect that Apple will follow the same heretofore successful development path for the iPad 2 as for the first iPad: integrate mostly off-the-shelf components, with some technology development of a new higher resolution display. The new higher resolution display will not quite have the dots/inch of the iP4 “Retina” display, since doing so would require a larger number of pixels than any graphics processor is currently capable of, even for desktops. I expect the display to retain the 4/3 aspect ratio of the current display, since that works well for browsing and document viewing, but with a resolution of 1600 x 1200 pixels compared to the current iPad 1024 x 768.
Apple recently introduced AirPlay to provide high definition video playback capability from iOS 4.2 devices to Apple TV via WiFi. This is currently limited to 720p (1280 x 720 pixels), so Apple will retain this limit for iPad 2 in order to keep compatible with Apple TV. Despite the fact that most of the competing Android tablets will feature 1080p (1920 x 1080 pixels) output through physical mini HDMI connectors, I don't expect Apple to move to 1080p Airplay capability until they're ready to offer 1080p content on iTunes. Since downloading 1080p video over the Internet is still prohibitively slow for most users, I don't expect this anytime next year.
Borrowing from the current iPhone 4, the iPad 2 will be equipped with dual front and back facing cameras, with the front facing camera supporting the Apple video chat utility FaceTime, which works only over the Internet via WiFi rather than over digital cellular networks such as GSM. I also don't expect this to change for iPad 2, even iP2's equipped with GSM. The back facing camera will have the same capabilities as the iPhone 4 with 5 megapixel still recording and 720p video recording. I know most people will not want to use their iPads as video cameras, but the carry-over from the iPhone 4 will be just too easy for Apple to resist. Also borrowing from the iPhone 4, the iPad 2 will get the iP4's nifty gyro sensors.
For the successor to the Apple A4 processor, I expect Apple to once again license a design from Samsung.
As I reported in The ARMs Race, Samsung recently announced a new 1 GHz dual core successor to its “Humming Bird” CPU core featured in the Apple A4. Although I have speculated at length about internal processor development efforts at Apple, I doubt that these efforts have borne fruit in time for the iPad 2. Using my standard rule of thumb that production of iPads needs to start about 6 months before they appear in store shelves, then for iPad 2 to be available in April 2011, iPad 2 would need to have started production in October. The newly announced (on September 7) Samsung processor design is more likely to have been ready to slap into the Apple SOC than any homegrown Apple processor. But the timing is a bit tight, even for the Samsung dual core design. In the case of the A4, Samsung had made the announcement of “Hummingbird” in July 2009, giving Samsung more time to ramp up production of the A4 in time for the estimated start of iPad production in October 2009.
Alternate iPad 2 Realities
The above summary is my current assessment of the most probable feature set of the new iPad 2. Other slightly less probable variations include:
A lower screen resolution such as 1280 x 960 pixels, but still keeping the 4/3 aspect ratio.
A smaller screen variant, say, with a 7 inch diagonal screen. This is well within the Apple practice of spinning off different sizes of the same basic product, as in the case of the iPod.
A single core, but possibly faster version of the A4, in case the new Samsung based processor isn't ready in time. Alternate iPad 2 realities I really don't expect:
iPad 2 using an SOC with an Apple designed CPU. Despite having acquired Intrinsity, which had a wealth of SOC and ARM design experience, I don't consider Bob Mansfield capable of piloting such a product, although he is nominally in charge as Senior Vice President of Hardware Engineering. The departure of Mark Papermaster in August 2010 at least delayed if not derailed any Apple CPU design effort (as opposed to SOC design).
Removable media for the iPad.
Apple hasn't equipped any iOS devices with removable media, because doing so would make it just too easy to defeat Apple's digital rights management approach for content and apps. That won't change for iPad 2.
A physical connector of high definition video such as mini HDMI. Now that Apple has committed to the AirPlay wireless video approach, they can only go forward, not backward.
A significant iOS upgrade. I believe that iOS 4.2 to some extent anticipates a processor upgrade in iPad 2, so there's really no need (in Apple's view) for further significant enhancements to iOS in order to take advantage of the processor power in the new iPad 2.
The Competitive Environment
In almost every way that counts, iPad 2 will be a far better product than what Apple introduced this year, but it will also have another thing the iPad lacked: tough competition. By the time iPad 2 arrives in early Spring, there will be numerous competitors featuring dual core CPUs based on the ARM Cortex-A9 design being used in the recently announced Samsung processor and (by my presumption) in the iPad 2 SOC. One such competitor by Motorola was recently leaked and features a 10 inch touch screen similar to the iPad and runs the Android Honeycomb OS (that's next in line after the current Froyo and its successor, Gingerbread). The Mototab features an nVidia Tegra 2 dual core Cortex-A9 processor at 1 GHz, 10 inch 1280 x 800 display, microSD expansion, micro USB, mini HDMI HD video connection, and front and rear cameras. Other competition will come from Windows Tablet devices using Moorestown Atom processors. These processors will eventually feature multi-core and 64 bit processor variants, although it's not clear what will be available in early 2011. Even without multi-core and 64 bit computing, there's every reason to believe that they will be much more powerful computationally than the ARM based processors. Comparative performance testing has already demonstrated that the current generation Intel Atoms are much faster than ARM processors. I expect Moorestown to keep the performance lead over ARM while closing or narrowing the power efficiency gap. Windows Tablet machines won't be as light or power efficient as iPads or other ARM processor based Android tablets, but they'll also be much more like full-fledged laptops.
The basic problem for iPad 2 is that unlike iPad, it won't be really unique. Although iPad 2 will need a technology discriminator to keep it ahead of its competition, by virtue of Apple's risk-averse decision making, iPad 2 will offer merely comparable performance and features. Here, Apple's approach to product and technology development may prove fateful.
For some time, industry observers and analysts have been wondering what Apple planned or should do with the huge pile of cash (over US $ 50 B) it has amassed. At the time of the iPad introduction in January 2010, I considered the course to have been reasonably well articulated by Tim Cook that Apple was going to do processor development, and that the Apple A4 in the iPad was the fruit of that development. It turned out that the A4 was really more of the usual Apple exercise in systems integration, but at least it was a start. The hiring of Mark Papermaster, IBM's Power PC design guru, the purchase of PA Semi and Intrinsity, also appeared to underscore Apple's seriousness on the processor design front. However, the departure of Papermaster and his replacement by Bob Mansfield do just the opposite. But the most telling evidence that Apple has given up on processor design is the fact that the cash surplus keeps growing every quarter. Building a truly competitive processor design capability would require spending some significant portion of that cash surplus, which Apple just isn't doing.
Given the way Android phones are expanding market share and poised to overtake iPhone in the US, the handwriting should be on the wall for all to see, even the execs at Apple. Apple badly needs a technology discriminator, besides their wonderful operating systems, and processors are still the best target of opportunity. Yes, this is TD at its riskiest, with no guarantee of success. But doing nothing practically guarantees eventual failure.