Introduction
When I first learned of Atym, a flood of painful memories came darting through my mind. Throughout my career, I had experienced the many different embedded software issues the Atym team described - issues that caused significant project delays/cancellations, lost customer opportunities, and partner integration failures at both my customers and some of the semiconductor, embedded, sensor, IoT, and ML/AI companies I have worked for.
Fortunately, those painful flashbacks were quickly replaced with excitement as I learned how the Atym device edge orchestration solution addresses these challenges. I realized Atym leverages cloud-native principals from the likes of Docker and Kubernetes to enable developers to easily build, deploy, manage, and secure containerized applications for the resource-constrained devices used in embedded, IoT, and edge systems.
A Whole New World for Containers
While Docker, Kubernetes, and other technologies have allowed some edge/IoT systems to utilize containers, there are barriers to these implementations reaching resource-constrained MCU/MPU-based devices that have limited memory or cannot run Linux. The epiphany for me was when I learned that Atym’s powerful container solution had a memory footprint of just 256KB – 1,000x smaller than the typical baseline footprint of Docker on Linux! This allows Atym to enable and orchestrate containers for billions of resource-constrained devices that other container technologies could not support beyond these memory and/or Linux barriers.
In addition to enabling app containers all the way down to MCU-based devices, this tiny Atym footprint offers developers the ability to free up at least 256MB of system memory on an edge/IoT device compared to Docker and other embedded container technologies. Think of it as Docker on a diet! This extra memory capacity enables developers to add more software capabilities to a device or save on hardware cost by decreasing total system memory required to perform the same tasks.
Atym’s sweet spot is bringing cloud-like app containerization and orchestration to embedded, IoT, and edge devices that have between 1MB to 1GB of memory. This opens the door for a streamlined development process, greater innovation, and simplified field operations in myriad use cases and industries.
With Atym, an industrial sensor OEM now achieves flexibility to mix and match tiny app containers to perform different functions. Developers of software-defined networking devices now gain more memory capacity to add new features compared to using Docker. Makers of consumer electronic devices can now deploy containerized ML/AI models and update them on the fly without replacing or impacting the entire device image.
Relief for Embedded Software Challenges
Now that I have joined the Atym team as Head of Sales, I am eager to bring customers this solution to help address those embedded software issues I have witnessed causing so many challenges for organizations that could not utilize containers previously. Specifically, the following are five ways the Atym technology resolves key software challenges I have seen throughout my career:
Greater Development Flexibility and Efficiency
With Atym, software functions/applications for devices are developed and deployed as containerized modules, eliminating the headaches and inevitable project delays of integrating all the disparate source code into monolithic firmware.
Developers are able to program their specific functions/applications in the language of their choice (e.g. C, Go, Rust, Python), so they no longer need to have experience with specific coding languages. Each container on a device can also be programmed in a different language.
The Atym Hub enables CI/CD workflows and greatly improves the flexibility, portability, reliability, and maintainability of code by extending to constrained edge devices the same benefits that are taken for granted in the data center today.
Abstraction of Hardware Complexity
The Atym Device Runtime abstracts the lower-level complexity of embedded hardware into simple API calls to access system resources. Specific embedded HW knowledge is no longer required to develop applications on top of the Atym Runtime.
Applications in Atym containers are more readily portable across different HW (MCU/MPU, IO, components, etc.) and various products that use common HW.
Simplified Collaboration with Partners
Atym containerization eliminates the challenging task of integrating partner source code on a device (which I have seen cause months of delays). Developers can choose best-of-breed from different partners and simplify the integrations of various applications through containers (even when based on different programming languages).
Developers can quickly deploy partner code updates without affecting any other software running on devices or having to recompile the entire firmware image.
Valuable code IP is protected because integration can be done with obfuscated binaries instead of needing to share raw source code for compilation into a monolithic image. This applies to both internal and partner code and has become even more imperative in providing ML and AI models for system integration.
Streamlined Field Operations at Scale
Atym enables individual application containers to be updated without rebooting a device or impacting the overall code base, eliminating the risk of making a deployed device unresponsive. This is compared to a traditional monolithic firmware approach that requires the entire device to be re-flashed for any code changes which risks “bricking” the device. Updates can be organized into scheduled campaigns across large fleets of devices.
Only portions of code that need to be changed are delivered over the air, reducing consumption of transmitted data. This is key for edge devices that are deployed in remote sites.
Developers benefit from a number of remote diagnostic and troubleshooting capabilities for deployed devices and app containers including remote monitoring, log centralization, and remote restarts. This is performed through the Atym Orchestration Hub which can be on premise or in the cloud.
Security is improved through isolation between applications and the underlying HW and ensuring only trusted applications can be executed. Containers are only granted access to HW resources if the application has the required permission. Any application that has a security vulnerability can be quarantined without taking down the entire device.
Protection Against Proprietary Lock-In
The Atym solution is built on open standards including WebAssembly and the Zephyr RTOS.
The Atym Device Runtime is powered by the open source Ocre project. Atym seeded Ocre in the Linux Foundation’s LF Edge organization to provide vendor-neutral governance for ongoing development by the community. The Atym team is committed to encouraging community-driven innovation while enhancing trust in the technology so customers can avoid unnecessary lock-in and create valuable products.
Atym is also active in other industry efforts to facilitate Ocre becoming a de-facto standard for lightweight containerization. Examples include a new Special Interest Group (SIG) in the Bytecode Alliance focused on WebAssembly for Embedded/Industrial applications and the Linux Foundation's Margo industrial edge orchestration interoperability initiative.
Conclusion
It has been amazing to see how the Atym team has developed a tiny footprint container technology and powerful orchestration solution tailored for embedded, IoT, and edge systems to deliver a revolutionary change in the way embedded software can be developed and managed for billions of resource-constrained devices.
Many of the painful and costly limitations of traditional embedded development can now be addressed by leveraging the benefits of application containers that cloud developers have enjoyed for years. Key benefits include reducing engineering friction, accelerating innovation, and simplifying the manageability of device fleet operations. This provides organizations with significant time and cost savings, in addition to improving the experience of their end customers.
Additional Atym benefits include hardened security posture, better IP protection, improved ML/AI model deployments, and reduced memory consumption compared to containerization technologies like Docker. These benefits further organizations’ ability to greatly reduce project delays/cancellations, lost customer opportunities, and partner integration failures.
I am excited to see all the new and innovative solutions our customers can now deploy with Atym! Please don’t hesitate to reach out to learn more about Atym and how we might help.
Comments