This article was originally published on the O'Reilly Radar.
The nature of Apple's new iCloud service, announced at WWDC, is perhaps more interesting than it seems. It hints very firmly at the company's longer-term strategy; a strategy that doesn't involve the web.
Apple will join Google and Amazon as a major player in cloud computing. The 200 million iTunes users Apple brings with them puts the company on the same level as those other platforms. Despite that, the three companies obviously see the cloud in very different ways, and as a result have very different strategies.
Amazon is the odd man out. Their cloud offering is bare metal, contrasting sharply with Google, and now Apple's, document-based model. To be fair, Amazon's target market is very different, with their focus on service providers. If you're a Valley start-up looking for storage and servers, you need look no further than Amazon's Web Services platform.
Google and Apple's document model contrasts sharply with Amazon's service-stack approach. Both Google and Apple have attempted to abstract away things, like the file system, which stand between the end user and their data. An unsurprising difference perhaps, Google and Apple are consumer-facing companies that are marketing to the final end user rather than the people and companies who aim to provide services for those users.
But that's where the similarity between Google and Apple breaks down. Google sees the cloud as a way to deprecate general purpose computers in the hands of their users. In the same way that their new Chromium OS is built for the web, their cloud strategy is an attempt to move Google's users away from native applications so that their applications and data live in Google's cloud of services. Perhaps coincidentally, this also gives Google the chance to display and target their advertising even more cleverly.
Apple's approach is almost entirely the opposite. They see the cloud as a way to keep the general purpose computer on life support for a few more years until touch-based hardware is really ready to take over. Apple's new cloud platform is built for native applications, in an attempt to pull users into native apps designed for their platforms. This method also gives Apple the chance to sell hardware, applications, and content that will lock users into their platform even more firmly. This is the basis of the often remarked "halo effect."
At least on the surface things seem to be simple — the "why" of the thing is not in question. However it's what hasn't been said, at least openly, that raises the most interesting questions.
Apple is fundamentally platform orientated. It's deep in their company genetics. The ill-fated official cloning program from the mid-'90s, which was brought to a screeching halt by the return of Steve Jobs, seems to have set a deep fear inside the company about letting someone else control anything that might stand between the company and direct access to their customers.
At least to me, nothing confirms that mindset more than Apple's return to designing their own processors in-house in Cupertino. Apple has a long history of using its own custom silicon, but it's been more than five years since Apple has done so. With the move to Intel, the hope was to delegate nearly all of Apple's custom chip development. Unfortunately, that proved to be a stumbling block when Apple built the first generation iPhone. The Samsung H1 processor in the original model wasn't quite what Apple wanted, even though it was what had been asked for, and I think the return to custom silicon probably brought a sigh of relief in some corners of the company.
The link between custom chips and the cloud may seem tenuous at first glance, but I think Apple's return to designing their own silicon is telling. Almost as telling as spending half a billion dollars on a custom data center to support their new iCloud service. Both moves show the company is now committed more than ever to controlling the verticals. From the chips inside the devices to the data centers their customers' data ultimately resides on, Apple is committed to controlling the user experience, and the web has no place in that.
You might argue that this is because the web is "too open" and that threatens Apple's platform. However, the continuing argument over openness, or lack there of, isn't really relevant. Despite Google's protestations to the contrary, neither of these two companies is particularly open. The very document-based model they're both advocating in their cloud architectures precludes a truly open system. It's such an obvious straw man argument that it's not actually that interesting.
What is interesting is that there was little or no mention of the web, or HTML5, during Apple's
WWDC keynote. I think you'll see far less emphasis on HTML5 from Apple in the future, unless someone asks to do something with Apple's platform the company disapproves of, and then the traditional answer of "Well, you can
always do that in HTML5" will be rolled out again.
Apple has finally put their cards on the table. They have not yet bet the company on iCloud, but it's telling how deep the integration into both iOS and OS X appears to be. They have for too much invested in iCloud for it to fail, if only in reputation. Whether the first incarnation lives up to its promises out of the box is still to be seen, but success isn't out of the question. Despite MobileMe, Apple does know how to build large-scale reliable backend services. You only have to look at the App Store itself for an example.
So in the future don't be too surprised to see Apple integrate iCloud even more tightly with both iOS and OS X. For the same strategic reasons, don't be shocked to see more custom chips appear — I expect to see the arrival of ARM-based MacBooks and the transition away from Intel for Apple's laptops. That's because for Apple, It's all about the platform.
