Blog Post

Android Steps Closer to Fragmentation

Wind River announced today a commercial version of the Android platform that comes with pre-integrated apps and global support and is optimized for Texas Instruments’ OMAP 3. But the offering represents one more step toward a dangerously fragmented Android universe.

The product is targeted at device manufacturers as a faster, more efficient alternative to building a platform from the ground up. It is a validated, fully compliant OS based on the latest versions of the Android SDK, Wind River said. It comes with software from partners including Adobe (s adbe), PacketVideo and Red Bend Software, and features enhanced Android user interface and personalization options that enable carriers and manufacturers to present their brands on-screen.

Wind River’s platform underscores the appeal of open-source OSes, which allow developers to access the source code and tweak it as they see fit based on a variety of factors. But as James noted back in May, that flexibility is a double-edged sword, potentially giving birth to a host of flavors that look and feel differently from handset to handset and carrier to carrier. That means we could see Android updates that aren’t immediately available to all phones using the OS, forcing manufacturers to choose between modifying the updates for every phone or leaving some handsets without an upgrade entirely. And that could be a huge drag on a mobile OS that is enjoying an impressive amount of momentum.

15 Responses to “Android Steps Closer to Fragmentation”

  1. I would like to add to douglasawh’s point. This should reduce fragmentation, not add to it.
    A company like WR is leveraging their platform supplier role across many OEM’s/operators/suppliers, by fixing+hardening the common Android layers and offloading that effort.
    Wind River was the first “OHA commercialization partner”, right? I’m sure this is what Google had in mind. A relatively small number of qualified integrators helping to make Android work well on various hardware platforms, so that the OEM’s and carriers can differentiate without fragmenting.

  2. Colin Gibbs

    @ douglasawh — You’re right that Wind River’s product isn’t, in itself, a fork. But it makes it that much easier for ODMs to create their own versions of Android, which I think takes us one more step in the direction of splintering the OS.

    @ Gearhead Gal — I think it could be great for Android’s long-term prospects if the OHA could settle on a common middleware layer. But with 11 ODM members of the alliance (and counting), I think there are too many players vying for space for a true standardized layer to emerge.

    • Anonymous

      Standardization should be driven by the volume of ODMs, which is precisely my point. Linux operates at its core through commercialization partners that make it easier for solution providers to focus on what they do best. The marrying of the apps to the OS through the WR platform should continue to enable the open development for apps, unique user experiences and device components, like radios and screens and battery life. What WR can bring to the table is the ability to deeply integrate key enablers – like flash, video codecs, DRM, and graphics processors – into the Android stack so more vendors don’t have to re-invent the wheel for each Android device. A vendor can still pick and choose what to support in their go-to-market product. But the WR toolkit might make it easier for them to get products to market faster.

  3. The conundrum of openness is that while you can do ANYTHING, there is a support cost lifecycle associated with all of this flexibility.

    Samsung rolls out a set of phones with some proprietary hardware form-factor or feature set. Motorola releases an API for their proprietary social networking middleware layer so developers can build MotoBlur extensions. Verizon creates their own carrier-specific secret sauce, and rolls out their own proprietary Android Market.

    Now, as a developer, do you dive ‘all in’ and pay the recurring tweak and debug tax? Do you program to a lowest common denominator so as to get the widest device footprint coverage? Do you optimize to whatever Android device is the biggest hit, and ignore rest?

    The RIGHT answer to these questions is unknown, but history suggests that managing same with all of the resident self-interest involved across developers, handset makers, carriers and consumers, is REALLY hard, something that I blogged about in:

    Google Android: Inevitability, the Dawn of Mobile, and the Missing Leg

    Check it out, if interested.


  4. I’m not sure I see this as the omen you suggest. Perhaps the open source community needs a commercialization partner like Wind River to standardize across device manufacturers. Rather than seeing this as the evidence of fragmentation, what if the alliance of open handset manufacturers actually could find a common platform for marrying software innovations to hardware innovations? A middleware layer between the kernel and the custom apps which is built by a company that can support multiple CE vendors and phone manufacturers may actually provide a path to manage or control fragmentation. Wind River’s history of providing this service to CE manufacturers around Linux may position them to effectively address the issues.

  5. Great post. This fragmentation is a real problem for another reason: it dramatically complicates the work of app developers. The more OS flavors and phone form factors out there, the more expensive app development becomes. QA becomes a real nightmare as every change in code must be tested on all OS flavors and phones. That is why Apple is probably going to win this war. One OS, two form factors (ipod touch/iphone), a signal huge marketplace (itunes), and an easy way to get paid. This will lead to Apple always having more and better apps. When it comes to hardware, the platform with the most software wins.