Mark Asselstine
github profile: https://github.com/masselstine
Mark has worked in the software industry for 27 years, with over 24 years working on software used in critical infrastructure. He started his career designing and writing software in the fields of robotics, air traffic control and telecommunications. He joined Wind River in 2007, spending most of this time working on the Wind River Linux product and the open-source Yocto Project, while contributing and consulting on many Wind River products which include a Linux and/or open-source component to them. He has been the Principle Architect for Wind River's Open-Source Software Group for the past 5 years, where among other things he took over as champion of the "open-source first" approach for product development. Mark was also involved with proposing and executing a move of the open-source StarlingX project from a CentOS to Debian. Having spent many years working on commercial supported open-source offerings, he works hard to ensure a healthy balance between commercial and community interests. Since his first open-source software contribution in the late 1990s, his goal has been to be involved in the proliferation of open-source software. He is happy to have been around to see and contribute in open-source software's spread to (almost) everywhere.
Accepted Talks:
A unified approach for intelligent deployments at the edge
Over the past decade, ‘build from source’ solutions like the Yocto Project and Buildroot have been favored for enabling various use cases at the Intelligent Edge. Traditional methods of building embedded Linux devices, which offer extensive customizations and the ability to generate an SDK providing a cross-development toolchain, have allowed developers to maximize the performance of resource-constrained devices while offloading build tasks to more powerful machines.
However, the increasing connectivity demands of edge deployments, including Over-The Air (OTA) updates and new paradigms such as data aggregation, edge processing, predictive maintenance, and various machine learning features, necessitate a different architectural approach for both near-edge devices and servers. This results in using multiple distributions, creating a heterogeneous landscape of operating environments increasing complexity and cost. Such complexities impose a significant burden to monitor for CVEs and bugs, additional SBOMs, diverse update cadences, and many other challenges.
To address these challenges and provide a more homogeneous solution, this talk will discuss how we are creating a Debian derivative using modern tools, to ease maintenance while combining traditional installers with a new set of distro-to-order tools allowing a single distribution to better service edge and server deployment. Coupled with a unified tech stack, this initiative offers a strategic advantage for enterprises aiming to optimize their edge deployments, create a seamless operating environment across devices, and set the foundation for future innovations in intelligent edge deployments.