Container technology has always been part of the cloud domain, and as such, its roadmap has usually been dictated by the use cases and requirements of that world. In the servers’ domain, resource utilization is nowhere near as relevant as it is in the embedded domain. The different languages and technologies that power the tools and mechanisms through which containers are leveraged in the bare metal server and /cloud worlds just don’t fit into the requirements of embedded.
Despite the above, these past couple of years have seen an aggressive push from cloud-centric companies trying to tell the Embedded Linux ecosystem and its development community that we should make do with Golang, NodeJS and similar solutions and tools. Most are unaware of the challenges when you cram cloud tools into a resource constrained embedded system. Even though the architecture of some of these frameworks have the right intention (LXD), most just lack the interest in understanding the specific requirements of embedded.
In this talk, we’ll explore how using containers for embedded systems modelling can help facilitate development cycles by enabling modular software architectures. We’ll deep dive into what the real requirements of embedded systems are and how modern container technology can help us meet the actual needs of this world. And lastly, we’ll walk through an example with Pantavisor, an open source container framework implemented for embedded systems.