IoT Goes Open Source With the Linux Foundation’s Zephyr Project

Linux is well-known for being the OS of choice by adept PC users. It is also widely deployed in Internet servers and industrial control systems due to its small footprint, operating efficiency and open-source legacy. Attributes such as these also make it a natural choice to be a driving technology in the nascent Internet of Things or IoT.

The Linux Foundation recently announced the Zephyr Project, which is squarely aimed at bringing the best of Linux to IoT devices in the form of a small-scale real-time OS or RTOS. The project is being supported by some big IoT market players including UbiquiOS, Synopsys, NXP Semiconductors and Intel.

Zephyr as an IoT Accelerant

One of the many obstacles hindering IoT growth is the lack of interoperability standards that can be utilized by IoT device manufacturers. Zephyr could be the answer to that as a single, unifying, highly customizable, scalable and secure OS if the project is successful at meeting the extremely tight constraints required of IoT devices.

Linux already has many “micro” distributions that are used in a variety of applications as a reliable RTOS that uses a minimal memory footprint. Zephyr proposes to take these features to their limit in the nano-OS world of IoT.

The fact that Linux is open source, i.e. free, is an additional incentive to its use. Its adoption could go a long way toward realizing full value from IoT as it drops device costs and communication barriers between products from disparate vendors. Furthermore, being open source enhances the security of Linux, since code is open for scrutiny by security experts. Security is one of the biggest concerns around IoT, which is why Zephyr will include a dedicated security group.

How Low Can You Go?

There are at least a dozen small footprint Linux distributions made for PC users. One of the smallest is Tiny Core Linux [http://www.tinycorelinux.net/] with a six megabyte image that runs from removable media such as a USB pen drive. It even includes a GUI desktop. Others have taken Linux “distros” down to 200KB of RAM for storage and one megabyte of flash for OS execution.

Such a minimal footprint is not small enough for IoT devices controlled by relatively simple microcontrollers that use no more than 10KB to 100KB of memory for storage and OS. Furthermore, in that sparse space, an IoT device is expected to operate in real-time. These constraints are formidable obstacles. The highly collaborative, community-based effort of Zephyr is an ideal approach to solving such problems.

Zephyr Goals

Zephyr is dedicated to producing both micro-kernel and nano-kernel versions. These are initially to be based on the contribution from Wind River Systems of its Rocket RTOS kernel. The nano-kernel project aims to run in no more than 10KB of RAM using a 32-bit processor.

More specific goals are stated for Zephyr as well:

  • It will be independent of CPU architecture, although it initially supports only x86, ARM and ARC.
  • Supported platforms include Arduino 101, Arduino Due, Intel Galileo Gen 2, MinnowBoard MAX, Quark D2000, Quark SE and the NXP FRDM-K64F Freedom board.
  • It will be modular, scalable and highly secure.
  • Connectivity support includes Bluetooth, Bluetooth LE, IEEE 802.15.4, CoAP, IPV6, and NFC.
  • Powerful development tools are part of the package with the kernel and select components open source under the Apache v2.0 License.

Long-term Impact of Zephyr on IoT Development

Despite IoT’s small-scale granularity, extreme scalability and potential for millions of new apps, its lack of standards and plethora of proprietary solutions have posed a significant hindrance to it being a hotbed for innovation. Diminishing such barriers, therefore, may be the Zephyr project’s greatest contribution to the growth of IoT.

By definition, open source platforms enable thousands of developers worldwide to invent new use cases, new apps and unique solutions. If the platform is inadequate for them to implement their solutions, they are free to modify the platform itself. This aspect will provide an enormous stimulant to IoT expansion. Zephyr’s primary goals to create an OS that is extremely flexible and scalable fits hand-in-hand with the open source development paradigm.

Conclusion

The Zephyr project will be seen in hindsight as a major inflection point in the progress of IoT. It is a boon to developers, enterprises and ultimately for consumers who will enjoy a much higher degree of interoperability, reliability and security among the IoT devices that inhabit their daily lives. Given that open source Linux is already utilized in a majority of enterprise networks and computational infrastructure, Zephyr adoption should be relatively smooth sailing as compared to a fragmented world of one-off OS solutions.