The future of Android likely means the death of Android

Opinion: Android has a lot of problems, and Google knows this, and is already making plans for the death of Android.
Written by Adrian Kingsley-Hughes, Senior Contributing Editor

On the face of it, Android seems like a very successful platform. It's the operating system that powers 85 percent of the new smartphones shipped, and, along with iOS, has decisively crushed the competition.

But Android is also plagued by dogged problems that Google can't seem to solve. First off, there's the issue of fragmentation, which means that developers have to create software that will work reliably on many hundreds of different devices from dozens of manufacturers. Sure, developers could just focus on devices from the big players - and some, such as Salesforce, are already doing just that - but that defeats the purpose of having a platform in the first place.

See also: Which Android smartphones get updated the fastest and crash the least?

Then, there's the update problem. Google has a yearly release schedule in place for Android, but it takes the new version a good part of a year to break the 10-percent usage mark. And it takes about four years for a release to fully penetrate the ecosystem.

Four years is a crazy long time in tech.

While Google is speedy at getting new releases onto its Nexus hardware, it can take months for owners of Android devices from other manufacturers to start seeing the update. And many never see the update.

In fact, outside of buying a Nexus device, the easiest way to get your hands on a new Android release is to buy a new device.

And you wonder why Android smartphone sales are so buoyant.

Here's how to make rechargeable batteries last for years

So, dominant market share aside, it's clear that Android has suffered from some extremely severe problems.

Many of the problems facing Android come down to it being open source. Android is all about being open, which means that once Google has released a new version, the OEMs and carriers are free to tinker with it to their hearts' content. That then results in both the fragmentation (OEMs loading the code onto any and every device form factor they can think of), and the problems with updates (Google can't push Android direct to devices because heavens knows what modifications and tinkering have been done, both cosmetic and structural, to the code).

Another problem with Android is that it's based on Linux, and Linux is both old and plagued by legal issues.

The Linux kernel was never designed for smartphones and IoT devices, and yet here we are shoe-horning it onto these devices. Sure, the kernel's been furiously tweaked, but tweaking can only go so far, especially when you're trying to optimize power consumption or when a platform that can run in real-time is required (the Linux kernel isn't real-time and instead uses a scheduler).

Also, all that legacy code is the perfect breeding ground for bugs and vulnerabilities.

The legal issues, specifically intellectual property licensing issues, are a real thorn in the side for those making Android devices. Companies such as Microsoft pulling in billions of dollars in licensing fees from Android hardware OEMs, which eats into the already razor-thin profit margins.

A totally new, built from the ground-up platform could free OEMs from being shackled to expensive patent licensing deals.

And Google is working on new platforms. Take, for example, Project Fuchsia. Sure, it's early days, but it's clear that Google is looking to an era beyond the Linux kernel, and the end of the Linux kernel ultimately means the end of Android. This operating system not only could be built from the ground up so as to be optimized for today's devices, but it would also be free from IP headaches.

It could also be modular in nature allowing for it to be customized for different applications - from the desktop and laptops to smartphones and even small IoT devices. It would be a truly unified platform. Google could license this platform to hardware developers, as opposed to using the open source model.

Licensing would give Google greater control over the hardware that Android was loaded onto. No more junk that can barely run the OS. No more devices being launched running years-old releases with no hope of an update.

This would truly be a platform for the 21st century.

Don't expect this to happen anytime soon, as operating systems take time to develop. But equally, don't think that Android is going to be around forever. Android has some serious shortcomings that Google is well aware of, and the fact that it isn't making much effort to fix these issues is perhaps the strongest evidence - along with the company's declared interest in developing new platforms - that such work is ongoing.

How long until we see an Android replacement? I'd guesstimate something in the region of five years, but a company such as Google has the resources to make it happen much quicker. But it could also lose interest in the idea and decide that Android is good enough as it is.

A dozen Raspberry Pi alternatives

See also:

Editorial standards