Last week, a group of Google people did a strange thing: they quietly disclosed a new operating system, in theory, this system with Android their own OS Google is a competitive relationship.
Currently in research and development of this set of open source operating system code named Fuchsia, which can be run on everything, down to the light of the amount of single purpose devices (such as ATM and GPS units), up to the desktop computer can support. But unlike Android, Fuchsia is not based on Linux, nor is it derived from any other software that constitutes the basis of personal computing and communication. Instead, it's a try from scratch.
Fuchsia is still in the early stages of development, as to how the system intends to use, Google has not released any major announcement, this is only an experiment the possibility of the existence of. Despite this, Google still has a lot of reasons to reset decades of software history.
This has something to do with your cell phone, tablet and notebook, but you may not be aware of it: most of the software that they use is very old. Linux uses the Android kernel, which was developed in 1991. OS X Mac, iOS and other apple platforms are based on Unix, which is derived from the Baer laboratory in 1969. Windows computer based on the NT Windows kernel to be traced back to 1993.
The purpose of the kernel is to manage the lowest level of the operating system. It handles the request of the keyboard and other hardware devices, the task scheduling, and management of files and memory. To this end, it is necessary to abstract the complexity of the operating system, and this is helpful, for example, you can let the developer do not need to know the specific printer models to print.
For an industry obsessed with the latest technology, such as Unix, Linux and NT Windows and other old kernel flexibility seems to be a little contradictory. But industry analyst Dediu Horace believes that, from the bottom level of view, the calculation is basically the same as a few decades ago. For example, today's Windows computer using the chip is the first PC IBM direct descendant of the Intel processor. In this sense, the kernel is nothing more than a commodity.
"We still use the same architecture, still use the same computing concept -- register, a gate circuit, a transistor -- for this reason, and has nothing to do with a better kernel necessary," dediu says: "the kernel problem has been solved."
We probably think so, too. This period of time since we put the sensors and computation capabilities packaged into something more, for example, the daily family becomes a smart home, and usually all connected more closely (jargon called the Internet of things). The idea behind the Fuchsia might be that aging kernels like Linux are not enough to cope with the new wave of equipment. Similarly, the creator is also thinking of a new kernel for the modern. (the kernel itself is called Magenta, based on the latest LittleKernel Google).
Supalla Particle's Zach provides hardware tools and developer tools for the Internet of things. He pointed out that for these small scale computing devices, Linux there are a number of issues.
First, Linux for this kind of application is too big. Even if the Linux kernel is modular, allowing developers to remove unnecessary parts, but will eventually occupy a few trillion of space. This means that it is more difficult to put the Linux kernel into a cheap microcontroller inside, so that the processor becomes necessary, and the latter is often much larger, more expensive and more energy consumption.
"The new supply chain this gap has not crossed, which made the manufacture of higher quality requirements, and the cost is much lower," Supalla said.
Another problem is that Linux is not "real time". ATM, medical products and other single use devices are different from the embedded system, Linux use of scheduling mechanism to deal with a large number of tasks. Although this can greatly play the performance of general-purpose computers, but also led to the need for precise timing of the device, such as 3D printers or car internal many electric control system problems.
"If you want to make sure that this thing is running in the exact moment of time, you don't want to have a process where you decide when to run," Supalla said.
Supalla said that for the Internet of things applications, such as Linux General operating system security may also be lower. Its code is larger, which means there may be more security vulnerabilities need to be processed or locked by a firewall or VPN.
"One of the value of running a real-time operating system or embedded system is that the system doesn't have to lock anything," Supalla says. "It doesn't have to run a lot of things you need to think about. Except for the software you write, it doesn't run anything."
Supalla inference, fuchsia is set two major attempt, in let applications and hardware communication on Linux via the operating system still handle better, and some of today's embedded operating system, such as FreeRTOS and ThreadX there is no those problems mentioned in the Linux.
"They may want both Linux that level of abstraction, wants to obtain a RTOS that performance, small size and real-time," supalla said: "these are very valuable things, and I think from the theory is achievable. I just didn't do it before."
If Fuschia's goal is just a small scale, it might not be worth the attention. But Fuschia's developers are more ambitious than this, and they claim that the operating system can be extended to smart phones and desktop computers. In theory, this would make Google a direct alternative to Android's Fuschia and OS Chrome.
Why do you want to do this? According to Supalla, it is possible to start from scratch to make a more efficient operating system, which in turn supports more efficient servers - Google has been very interested in this. He also pointed out that desktop compatibility can also simulate a large number of smaller devices running, so as to ensure the scale of operation.
"Let one thousand servers at the same time running the same software applications, running more than one million chips, so it is better to test." Supalla says.
Dediu's theory is not the same: a new operating system allows the Google to stay away from the Android's intellectual property licensing problem. "Because this is a piece of white paper design, won't encounter someone to find IP (intellectual property) of trouble," he said, "this may be a reasonable assumption, because Linux does have some tricky IP problems."
Remember, this may be academic. Fuchsia developers said they would complete the record and eventually released the operating system, but the way to go is still very long, but it is not clear Google will give its full support. The Android ecosystem is already huge (and is merging with Chromebooks). At the same time, Google continued expansion of Android a version to networking equipment. This version of the features of the embedded system called Brillo, go to hell the disadvantages of Linux, it is becoming a mature platform, and not just a simple operating system.
Also, Unix started just a volunteer project, did not get the recognition of the Baer laboratory; Torvalds Linux do Linux is just his hobby. Perhaps a few decades later, we will discuss the Fuchsia in the Google had not been paid much attention to the matter.